Windows

Build a great PC on a budget with these parts and procedures

A Vista-capable system with quality components at a reasonable price can be yours--if you're willing to build it. George Ou explains the parts you'll need and provides illustrated, step-by-step instructions to guide you successfully through the assembly process.

This article is also available as a PDF download and a TechRepublic photo gallery.

A lot of people don't realize that it's possible to build a dual-core workstation with good 2D graphics and even some decent 3D performance that's fully Vista capable for a reasonable price. In fact, you can do it for around $1,133.  That price includes a massive 22" LCD and Vista Home Premium OEM software, as well as the cost of shipping.  For that kind of money in the retail sector or mail order, you wouldn't get close to these specifications and you'd probably end up with lousy embedded graphics chipsets.  The only catch is that you actually have to build it. But in this tutorial, I'll explain the parts you need and show you how to assemble it all.  Not only do you get the satisfaction of knowing you've put in good components, you also get the satisfaction of giving the PC life with your own hands. I can't think of a better hobby to have.  We'll start with the parts list.  Note that the pictures shown aren't necessarily the same parts in the list, but they're close enough for the purposes of this tutorial.

 

Parts list

Mainstream home or office dual-core workstation Cost
Biostar TForce965PT with Realtek gigabit LAN and 7.1 audio * $105
Intel dual-core C2D E6300 (can clock very high) ** $184
Stock Intel retail CPU fan (included with CPU)  
Seagate 400 GB SATA II HDD $120
Cooler Master CAV-T03-UW $70
NVIDIA GeForce 7300 PCI-Express with 256 MB (w/HDTV out) $70
Patriot 1 GB (2 x 512MB) DDR2-667 RAM *** $72
SeaSonic silent/efficient 330 watt PSU $60
Lite-on 16x SATA-based dual-layer DVD burner $36
Dell 22" widescreen 1680x1050 LCD monitor $296
Vista Home Premium Edition OEM 32-bit (64-bit same price) $120
Total (shipping included but not taxes) $1,133

*You will need Vista drivers found on Realtek's Web site for LAN and audio.  This is especially true if you get Vista 64-bit edition, since the sound drivers included won't run correctly without the updated drivers.  All other hardware should run fine on a fresh Vista install, though it's recommended that you use the latest 32-bit NVIDIA drivers (64-bit).

** This CPU can typically safely overclock 25% over the stock speed with minor voltage increases to keep it stable in Vista and permit SpeedStep power saving mode to function.  Some more aggressive overclockers have been known to push this to 50% and even beyond, with high voltage increases and massive cooling fans.  Do not attempt more than 25% with the stock Intel CPU fan and stick with modest voltage increases if you want your system to be stable and without error.  You can read more on how this type of hardware works with Windows Vista.  This type of a mainstream system isn't meant for massive overclocking and a free 25% boost in performance (bumped up to 2.33 GHz) is very safe and doable.  Also note that you'll need to bump up to DDR2-800 if you want to clock the CPU any higher.

*** 1 GB in Vista will offer good performance.  If you double the RAM to 2 GBs, it will offer premium performance and allow you to work with much larger graphics files or run multiple virtual machines inside the free Virtual PC 2007.

 

Assembly instructions

We'll start by taking the PC chassis out of the box.  The chassis shown is the Cooler Master CAV-T03-UW, which is solidly built and relatively cheap at $60.  There's usually a small bag or box that contains screws you need to put the system together.  In Photo A, it's lying on the chassis on top of the drive bays.  I've also taken the power supply out of the box and laid it inside the chassis, shown in the upper-right of the photo  Note how the power supply has the fan grill exposed toward the motherboard. That is the orientation you want.

Photo A

Next, we need to find the following kind of screw to hold down the power supply shown in Photo B.

Photo B

Now, we need to use four of those screws to screw in the power supply.  Note the location of the four highlighting circles I drew in the bottom left of Photo C.

Note on the power supply: I'm using a SeaSonic S12-330 330 watt power supply, which can be had for $55.  I swear by these power supplies, and I own five of them because they're super efficient at above 80 percent, they're dead silent, and they're very affordable.  Most power supplies make much more noise and are typically in the 60-70 percent energy efficiency range.  Many people spend upwards of $100 on a higher power supply, citing the myth that you need at least 400 watts and higher on a modern computer.  That's utter nonsense even for a high-end PC. Even the highest-end PCs peak out at around 250 watts.  The measured power consumption at the plug for the PC in this tutorial idled at around 73 watts and peaked out at around 110 watts during intense CPU loads.  The high wattage power supply proponents will often cite video card manuals specifying that they need a 400 watt power supply, but those numbers were just pulled out of thin air and aren't based on actual power consumption measurements.  330 watts happens to be one of the smallest ATX power supplies you can find, and it's more than enough for what we're building here.

Photo C

Next, we have to find the following types of screws, shown in Photo D, to mount the motherboard in place.  You'll usually need nine of them.

Photo D

You can mostly put these screws in with your finger, but you'll still need something like the tool shown in Photo E.

Photo E

Note the location of the nine red circles shown in Photo F.  That's where the screws typically go for mainstream ATX motherboards.  Some motherboards may be smaller, so you'll need to mount the three top screws one notch lower.  Be sure to examine your motherboard to confirm the placement of the screws.  You'll need to use the tool shown in Photo E to tighten them, but do not make them so tight that you strip them.  It's just as bad if you make them too loose, because you'll never be able to get your motherboard out.

Photo F

Next, you'll need to find the I/O panel that came with your motherboard, which will look something like the one in Photo G.  Be sure to knock out the necessary holes if your motherboard has that component.  The red square shown in Photo G is the main LAN interface port hole, and it needs to be knocked out.

Photo G

Carefully knock out the old placeholder I/O plate with the dull end of a screwdriver. (If you use your finger, you could get cut.)  You want to strike from the outside in.  Then, place the new I/O plate in from the inside and press all four corners in place, as shown in Photo H.  I usually do this with the dull end of the screwdriver hitting all four corners so I don't get cut.  The edges here are sharp.

Photo H

Next, you want to find the kind of screws shown in Photo I to hold down the motherboard.  You'll need nine of them.

Photo I

Gently screw them in where the red circles are in Photo J.  DO NOT overtighten because if you lose a screw below the motherboard, you'll never get the screw out.  You want it all the way in but not too tight, unless you want a stuck motherboard.  The only way to get it out will be to take something and carefully cut the screw without damaging the motherboard--not something you want to have to do.  Note that for Intel-based Socket 775 systems, you should probably put the CPU and fan on the motherboard before you put the motherboard in, since those things aren't easy to put into place with the flexing of the motherboard.

Photo J

If you didn't do this before you put in the motherboard, find the CPU socket and remove the cover, shown in Photo K.

Photo K

Photo L shows an Intel Socket 775, used in most recent Intel-based desktop PCs.  You must open the lever and then pop open the lid to expose the pins in the socket.

Photo L

Photo M shows a retail Intel Core 2 Duo E6400 and the fan it comes with.  In the red square is the CPU.  The radiator device on top is the CPU cooler.

Photo M

Take the CPU out and remove the protection cap, as shown in Photo N.  Note the positions of the CPU notches. I've circled them because they have to be aligned with the CPU socket.

Photo N

Photo O shows the chip in place.  To get it there, you simply line up the notches and place the chip on the socket.  Close the lid and then close the lever and secure it under the hooks.

Photo O

Photo P shows some generic thermal grease.  A much better material made of liquid metal is available, and you might want to use that if you have a higher-end system.

Photo P

Photo Q shows the CPU cooler in place.  Putting in those four pins isn't easy because of the motherboard flexing away.  That's why it's a good idea to put the cooler in before the motherboard.  All four pins need to be snug and securely inserted into the motherboard holes.  Note the four-pin connector, shown by the red rectangle. You use that to power the fans.  The four-pin connector allows RPM readings as well as dynamic fan speed control based on the temperature.  The stock Intel CPU fan works well for all Intel processors when the chip runs at official stock speeds.

Photo Q

Photo R shows a mess of front LED and switch connectors going to the chassis.  It's a crying shame that in 20 years, the motherboard industry hasn't figured out a way to standardize this so that you have a simple connector for everything.  It hasn't really changed all that much.  Even the front speaker/microphone connectors are a mess, and I actually decided not to hook them up, since I have 5.1 audio anyway and the front ports on the chassis support only 2 channel audio.  Unfortunately, you'll have to dig out the motherboard manual and figure out where the pins for all these connectors go on the motherboard, and you might want to use some long tweezers if you have them.

Photo R

Photo S shows all the connectors for the front of the chassis plugged in.  Note that the HDD fan connector shown on top is for the 80 mm fan cooling the hard drives in the bay.  I generally don't use this if there is only one hard drive in the bay, since it can cool off relatively easily.  Two or more drives definitely need it. 

Photo S

Photo T shows the motherboard power connectors, with the SATA power cables in the red rectangle on the bottom.  The motherboard connectors are plugged in, but the SATA connectors aren't yet.  Be careful how you connect these motherboard connectors and make sure the clips line up and snap into place for both the left and right connectors shown in Photo T.  If you reverse it, or worse, shift it over (which is pretty hard to do), it's possible you'll see smoke coming out of the motherboard when you try to plug it in and turn it on. I've seen someone do this before, and it isn't pretty.  If nothing snaps into place and you can easily pull it out without having to squeeze on the release clips, it's plugged in wrong.

Photo T

Photo U shows a Gigabyte made NVIDIA 6600 PCI-Express 128 MB video card with completely silent passive cooling.  This particular video card seems to be discontinued, and a NVIDIA 7300 with 256 MB is the closest I could find in the $65 price range.  It's not for gaming, but it's perfect for business and non-3D graphics design usage.  It can be used for moderate 3D gaming, and it's much better than the embedded Intel or NVIDIA 6100 graphics chipsets you get with those expensive retail computers.  If you want a higher-end model with dual digital DVI ports, you can try an NVIDIA 7600GT card for $120 on an "open-box" model.  I must also stress the fact that all the cards I'm showing here are passively cooled, which means they contribute zero noise to the system.  The finished system shown in this tutorial is almost completely silent, which is music to my ears.

Photo U

Photo V shows a 400 GB SATA II hard drive.  To the left of the image are the SATA power and data connectors.  These drives can be obtained from discount retailers for around $110.

Photo V

Photo W shows a new Lite-on 16x SATA-based dual-layer DVD burner, which is now only $32.  DVD drives have always used those wide PATA connectors, but most new motherboards have only one--or often, no PATA connectors at all.  On the other hand, even a cheaper motherboard will have four SATA connectors.  This is the first time I've used an SATA DVD drive, and I will say it's a pleasure to work with compared to PATA.

Photo W

My drive came with a black faceplate, but the case is silver.  So I had to swap out the faceplate with the gray one that came with the drive in the box.  Photo X shows how you take these apart.  First, you have to use the pin (paperclip will work) and insert it in the hole in the front to pop open the tray.  Then, you slide the tray lid up and out and take off the front panel.  You put the new panel on and then the new tray lid.

Photo X

Photo Y shows the DVD burner and hard drive inserted into the chassis.  This particular chassis uses a screwless locking mechanism to hold the drives in place.  The photo also shows the PCI-Express Video card plugged in at the bottom-right of the photo.

Photo Y

Photo Z shows the gray SATA data cables connected and the SATA power cables plugged into the drives.

Photo Z

Photo AA shows the DDR2-800 memory DIMMs.  DDR2-533 is all that's needed for this system, but since the price isn't all that different, I went with the higher-end memory.  Note where the notches are because they need to line up with the motherboard memory socket.

Photo AA

Photo AB shows the memory DIMMs inserted.  Be sure you line up the notches and open the side clips before you insert the memory.  It should work its way in relatively easily, and you should feel the clips on the side snap into place.  Squeeze the clips in to make it secure.

Photo AB

Now, everything should be working. But before you put the side panel on, try to see whether the system BIOS will post.  If it doesn't, don't panic. It's usually something you forgot to connect or something that isn't seated properly.  Just check all your connectors and it should work.  If the LED lights in the front don't light up, simply reverse the connectors.  Those LED connectors are no big deal, and nothing bad will happen if you plug them in wrong.  Every other connector in the system is pretty much idiot-proof, since you'd have to be formidably strong to connect them incorrectly.  Just make sure the clips and notches are all aligned correctly.

188 comments
weaverita
weaverita

Nice post. I am looking to build a 2U version of this but have never found an inexpensive 2U frame / chassis that would accommodate 6 (or 8) hot sway drives. It seems that everyone that has the hot-swap bays up front only sells the 2U box as a complete solution that runs $Ks. Pointers welcome.

ojeda
ojeda

Very nice piece of work.

jck
jck

Got an itch to do this...hope my HTML works, so my list probably won't be as pretty, I don't think. My computer didn't want to do plain HTML. But, this is how I built a machine with a total cost of about $1250 including shipping and mail-in rebates. [b][u]jck's basic game machine design[/u][/b] Combo - EVGA 122-M2-NF59-TR Socket AM2 NVIDIA nForce 590 SLI MCPATX AMD Motherboard - Retail + AMD Athlon 64 X2 5600+ Windsor 2.8GHz Socket AM2 Processor Model ADA5600CZBOX - Retail, $311.40 G.SKILL 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory Model F2-6400CL5D-2GBNQ ? Retail,$104.99 RAIDMAX SMILODON ATX-612WBP Black SECC STEEL ATX Mid Tower Computer Case 500W Power Supply - Retail, $105.98 2x Seagate Barracuda 7200.10 (Perpendicular Recording) ST3250620AS 250GB 7200 RPM SATA 3.0Gb/s Hard Drive - OEM, $139.98 SAMSUNG Black 1.44MB 3.5" Internal Floppy Drive Windows 98SE/ ME/ 2000/ XP - OEM, $11.98 ARCTIC COOLING Freezer 64 Pro 92mm CPU Cooler ? Retail, $34.98 2x LG 18X DVD?R Super-Multi DVD Burner With 12X DVD-RAM Write Black IDE Model GSA-H42N-BK - OEM, $71.26 XFX PVT73GU1D4 GeForce 7600GT 256MB GDDR3 PCI Express x16 Fatal1ty Professional Series Video Card - Retail, $159.99 Hanns?G HW-191DPB Black 19" 5ms DVI Widescreen LCD Monitor - Retail, $183.46 Microsoft Windows XP Home With SP2B 1 Pack - OEM, $94.98 LITE-ON SK-1688U/B Black USB Wired Standard Keyboard - Retail, $12.13 inland 07233 Black 3 Buttons 1 x Wheel USB Wired Optical Mouse - Retail, $9.48 Logitech X-530 70 watts RMS 5.1 Black Speaker System - OEM, $74.36 Total (shipping included but not taxes) $1,314.97 Rebates (not including the ~$2 in stamps and envelopes) -$65.00 Grand Total: $1,249.97 Pros: price, added high-performance CPU cooler for OCing, doubly better video performance, estimated 20 percent processor performance increase, rebates help offset shipping costs, solid performer, ability to do CD-to-CD or DVD-to-DVD duplication, 25 percent more drive space for 16 percent more cost, mine has a keyboard and mouse :D Cons: slightly smaller monitor, case p.s. shows only 1 PCIe connector and would require new p.s. for SLi operation, extra $20 for PCI 802.11g card, might need SATA power splitter (about $6) You could remove the floppy drive and get a 512MB or 1GB flash drive instead...I'm just nostalgic for my floppies. You could also save $95 by using a free Linux distro and using Wine, or save $40 and get 1 year of Transgaming Cedega with your Linux All prices shown are what I was quote from the Newegg.com website as of 04/27/2007. Your prices may vary. Edit: HTML went out the window...

TechExec2
TechExec2

. George's statements about AMD CPUs quickly fall apart when you examine the "data" used to make them. [b]George's false statements about AMD[/b] George says [i]"...AMD chips consume a little more power in idle..."[/i] and [i]"...AMD chips consume a LOT more power at peak workloads..."[/i] (1) and cites testing by TheTechReport as "evidence" (2). [b]The TechReport "evidence" cited by George[/b] Note: This data will be discredited below. [pre] 65 nm CPUs Idle power Difference Clock (GHz) ------------------------------------------------------------------------ Core 2 Duo E6700 65 nm 120 +0.0% 2.66 Core 2 Duo E6400 65 nm 121 +0.8% 2.13 Core 2 Duo E6600 65 nm 121 +0.8% 2.40 Core 2 Extreme X6800 65 nm 122 +1.7% 2.93 Athlon 64 X2 5000+ 65 nm 123 +2.5% 2.60 Core 2 Duo E6300 65 nm 125 +4.2% 1.86 65 nm CPUs Peak power Difference Clock (GHz) ------------------------------------------------------------------------ Core 2 Duo E6400 65 nm 143 +0.0% 2.13 Core 2 Duo E6300 65 nm 153 +7.0% 1.86 Core 2 Duo E6600 65 nm 163 +14.0% 2.40 Core 2 Duo E6700 65 nm 165 +15.4% 2.66 Core 2 Extreme X6800 65 nm 172 +20.3% 2.93 Athlon 64 X2 5000+ 65 nm 187 +30.8% 2.60 [/pre] [b]Examining the "evidence"[/b] - I included all of the TechReport data regarding 65 nm CPUs. It's no news that 90 nm CPUs draw more power than 65 nm CPUs. - The TechReport data shows that the AMD-based [u]system[/u] draws 2.5% more power than the best Intel-based [u]system[/u] at idle and 30.8% more power at peak. - The TechReport data shows that worst Intel-based [u]system[/u] draws 4.2% more power than the best Intel-based [u]system[/u] at idle, and 1.6% more than the AMD [u]system[/u]. Oops! :^0 - The TechReport data shows that the worst Intel-based [u]system[/u] draws 20.3% more power than the best Intel-based [u]system[/u] at peak. Oops! :^0 [b]Conclusion[/b] George says [i]"...AMD chips consume a little more power in idle..."[/i] and [i]"...AMD chips consume a LOT more power at peak workloads..."[/i]. This is false. The TechReport data George cites also shows that [u]Intel chips consume a little more power than Intel chips in idle[/u] and [u]Intel chips consume a LOT more power than Intel chips at peak workloads[/u]. This is ridiculous. The TechReport testing was severely flawed (3). The TechReport data is discredited. George's statements about AMD chips are discredited. George's bias against AMD is exposed. This is obvious. George Ou doesn't need no stinkin' facts to run AMD into the ground. ------------------------------------------------- (1) AMD chips consume more power http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2220200 (2) TheTechReport.com http://techreport.com/reviews/2007q1/cpus/index.x?pg=13 (3) Interesting but severely flawed article http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2220261

TechExec2
TechExec2

. [b]AMD "Barcelona" Demonstration - November 2006[/b] Last November, AMD demonstrated "Barcelona", the first true quad-core X86 CPU (1). "Barcelona" is the only true quad-core X86 CPU ever demonstrated publicly. Intel's "quad core" CPU is really two dual-core CPUs in the same CPU package. It's a very nice CPU, but it's not really a quad-core CPU. [b]New details about "Barcelona" announced[/b] Yesterday, AMD released new information about "Barcelona" and it is very exciting!! Read all about it here: (2). The first version of "Barcelona" will appear in the Opteron server line. In addition to having four cores on a single piece of silicon, there are many architectural performance improvements including the following (if you're not into this stuff like I am, suffice to say that there is a lot more to like about "Barcelona" than twice as many cores): - SSE instruction execution width: Doubled - Instruction fetch bandwidth: Doubled - Data cache bandwidth: Doubled - L2/Northbridge bandwidth: Doubled - Floating point scheduler depth: Doubled - 64 KB on-chip L1 cache for each core - 512 KB on-chip L2 cache for each core (total 2 MB) - 2 MB shared on-chip [u]L3[/u] cache - Total on-chip cache: 4.25 MB - Improved branch predictor - Sideband Stack Optimizer - Load reordering (efficiency) - Larger translation lookaside buffers - Two separate independent 64-bit on-chip memory controllers - Other memory controller improvements - Data and instruction prefetcher improvements, 2 prefetchers per core (8 total) - New DRAM prefetcher in the memory controller with its own buffer - Virtualized address translation performance improvements - Power management improvements. Each core can throttle down independently if unused. - Same thermal envelope as current dual-core Opterons. Note: Intel "quad core" processors are really two dual-core CPUs and require DOUBLE the power of a single dual-core CPU. - 65 nm manufacturing process - Plus other improvements [b]AMD's answer to the Core 2 "Conroe"[/b] There will be two desktop CPU lines based on "Barcelona": A quad-core CPU named "Agena" (2.7 to 2.9 GHz) and a dual-core CPU named "Kuma" (2.0 to 2.9 GHz). [u]This is AMD's answer to Intel's Core 2 "Conroe" line[/u]. [b]Conclusion[/b] It is very good to see AMD's response to the Core 2 and it looks to be very strong! Of course, by the time these CPUs are available, the Core 2 will have been on the market for over a year. This is late. But, late is better than never. AMD still has game! And remember, so long as AMD keeps on providing great price and performance, everyone should still buy only from AMD right now as I explain here: (3). The competition must be preserved and AMD is still too much smaller than Intel. ----------------------------- (1) AMD offers quad-core tease http://www.eweek.com/article2/0,1895,2066276,00.asp (2) Barcelona Architecture: AMD on the Counter-Attack http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2939&p=1 (3) ALWAYS use and AMD CPU when you build your PCs - Here's why http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2180346

pickleman
pickleman

This is anything BUT a "budget PC". 22-inch LCD? 400 gig hard drive? Get real. A "budget PC" is something you can put together for less than $500, doesn't run bloated Vista, and sure as hell doesn't have a 400 gig drive or 22-inch LCD. It's obvious that your experience as a system integrator and/or solutions provider is considerably lacking..

TechExec2
TechExec2

. This article of yours is going to help a lot of people. Great job!

TechExec2
TechExec2

. Ford's fastest car, the Ford GT, is faster than Chevy's fastest car, the Corvette. Does that mean that all Fords are faster than all Chevys? Of course not. [u]Contrary to what the Intel fanboys say, all Intel CPUs are not faster than all AMD CPUs[/u]. Don't be fooled. It's simply untrue. For a bunch of years, AMD made the fastest desktop CPU. Right now it's Intel, and it costs about $1000. I never buy the top speed CPU no matter who makes it due to poor value. Below that however, [u]AMD has a price and performance match for every Intel CPU[/u]. This is not an accident. AMD and Intel both price their CPUs so they compete well with each other. This competition is very important! [b]Why do I care about AMD? Why should you?[/b] The answer is simple. Without the presence of AMD as a strong competitor, the cheapest Intel desktop CPU will cost $1000, not the most expensive one. I don't want to go back to those days, do you? Since AMD offers great price and performance on the CPUs I actually buy, I always buy AMD. I get the speed I want at the price I want, and [u]I get a FREE bonus: I help keep AMD in the game and keep Intel pricing in check![/u] Now, [u]that's[/u] important! :-) [b]AMD Fanboy? No.[/b] I'm not an AMD fanboy. This is about my best interest. I'll be buying Intel CPUs again when (1) AMD is about as financially strong as giant Intel, or if (2) AMD stops providing great price and performance for the CPUs I actually buy. Until then, it's AMD-only. [b]Call to Action[/b] [u]Buy AMD CPUs[/u]. It matters. Help keep AMD in the game. Keep Intel from becoming the only supplier of high-speed X86 CPUs. Keep Intel prices from rising through the roof. After all, it costs you nothing and it is in your best interest, wouldn't you agree? :-) When you support Intel by buying from them, [i]YOU DON'T WIN, INTEL DOES[/i]! Think before you act. Be a winner. Buy AMD. All the best to you! [b]Choose the AMD Dual Core CPU that meets your performance needs[/b] I prefer the PC WorldBench to compare CPUs. I think it best represents the performance I will actually see. Higher is better in (1) and lower is better in (2). [u]Processor.....PC WorldBench Score.....Price[/u] AMD Athlon 64 X2 4600+ (2.40 GHz).....124 (1).....$189 (3) Intel Core 2 Duo E6300 (1.86 GHz).....122 (1).....$183 (4) AMD Athlon 64 X2 5000+ (2.60 GHz).....130 (1).....$215 (3) Intel Core 2 Duo E6400 (2.13 GHz).....131 (1).....$221 (4) AMD Athlon 64 X2 5600+ (2.80 GHz).....138 (1).....$332 (3)(5) Intel Core 2 Duo E6600 (2.40 GHz).....140 (1).....$314 (4) AMD Athlon 64 X2 6000+ (3.00 GHz).....489 (2).....$540 (3) Intel Core 2 Duo E6700 (2.66 GHz).....482 (2).....$512 (4) Note that this Intel CPU fares better on the Photoshop test (2). AMD currently has no match for this performance level Intel Core 2 Extreme X6800 (2.93 GHz).....156 (1).....$965 (4) [b]AMD Partners Supply Great Motherboards[/b] There are many great motherboard suppliers for AMD CPUs including ABIT, ASUS, BIOSTAR, GIGABYTE, MSI, and others (6). Personally, I have been using MSI and like them. -------------------- (1) PC WorldBench 5 tests http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2802&p=5 (2) WorldBench 5.0 Office XP SP2 Tests http://www.hothardware.com/viewarticle.aspx?page=5&articleid=929 (3) AMD Athlon X2 CPU Prices http://www.zipzoomfly.com/jsp/ProductList.jsp?ThirdCategoryCode=110501&SortBy=A (4) Intel Core 2 Duo CPU Prices http://www.zipzoomfly.com/jsp/ProductList.jsp?ThirdCategoryCode=110514&SortBy=A (5) AMD Athlon 64 X2 5600+ 2.8 GHz is equivalent to the AMD Athlon FX-62 2.8 GHz tested in (1). (6) Motherboards for AMD Socket AM2 CPUs (those I listed here) http://www.zipzoomfly.com/jsp/ProductList.jsp?ThirdCategoryCode=111211

Fil0403
Fil0403

I am not going to buy a worse CPU just for pitty or to preserve competition. If they want to be competition to Intel they have to release as good CPU's as them or better (which, at least currently, they don't), that's the law of every market, survival of the strongest.

w2ktechman
w2ktechman

it is not the cheapest PC, but it looks like a decent quality with extras for a pretty low price. So what is your problem with it? It's not $500 or less, I can build something for around that price, but George's system will eat it up. Plus George included Vista on it. He was putting together a full-functional Vista system at a low cost, to run all Vista apps and other programs. Your $500 would be slow at best on that. And if you want to bring Linux into it, that was not in the specs for this system, so it was just a slam for no reason.

TechExec2
TechExec2

. George's PC is a "GREAT PC on a budget", not a "Budget PC". There is a big difference between these two. And, it all depends on how big your budget is. Even people spending $1100 have a budget, it's just bigger than a small budget. It's people like you who give the word "budget" a bad name! :^0 Here's to wishing you win the next lottery...

jnowicki
jnowicki

I just wanted to say that I've used every mobo on the market to include ASUS, MSI, Gigabyte, Winfast and others. The best that I've used and the cheapest is ECS. GREAT mobos.

JohnnySacks
JohnnySacks

For me to endorse Intel with my wallet based on some irrelevant FUD issues equates to my wholehearted support for a total monopoly similar to the one Microsoft has in the OS and office suite market. The difference in this case is that the two competing products are identical (overkill) for any end use I will ever manage to throw their way. Should Intel rise as the single supplier of CPU's, we the consumers lose. I've used AMD exclusively since the K6 CPU's. They've ALWAYS been cheaper and I never noticed any performance issues whatsoever.

TechExec2
TechExec2

. Thanks for posting this, JCK. This test looks to be done a lot better than the one done by TheTechReport. For example, there are no obvious inconsistencies such as a faster processor running at a higher clock speed using less power. There are no "simulations" of real processor models. But, there is still a big problem: [u]Sharky did not use the new low-power 65 nm AMD CPUs[/u]! [b]My observations and comments about Sharky's tests[/b] 1. Sharky used actual CPUs clocked to spec, not "simulations" like TheTechReport. Closely matched AMD and Intel hardware. The only differences were the CPUs and motherboards -- as good as it gets. Excellent. 2. All of the Intel Core 2 CPUs tested, are 65 nm, and are rated at 65 watts. 3. All of the AMD CPUs tested were [u]not[/u] the lower power 65 nm "Brisbane" 65 watt CPU models. 4. The AMD CPUs tested were the 90 nm "Windsor" models. You can tell this by the model numbers and clock speeds except for the 5000+ model. The power usage of the 5000+ tested indicates the 90 nm model. 5. Desktop CPUs spend most of their time at idle. I still say that peak power usage is not the most important criteria for choosing a desktop CPU. [b]Power usage of AMD 65 nm "Brisbane" CPUs up to 11% lower[/b] According to this (1) article, the AMD 65 nm "Brisbane" CPUs offer up to 11% reduced power consumption at 100% full load when compared with their 90 nm "Windsor" counterparts. There's no question that the 90 nm AMD Athlon 64 X2 6000+ and FX-62 CPUs draw a lot of power (they are rated at 125 watts). But, we could apply Toms Hardware's -11% power savings to the other AMD test results from Sharky to approximate what the 65 nm CPUs would look like. This shows that the low-power 65 nm AMD CPUs are quite close to the Intel Core 2 CPUs when running at full load, just as you would expect (all of these 65 nm CPUs are rated at 65 watts). -------------------------- (1) AMD 65 nm "Brisbane" vs. AMD 90 nm "Windsor" Power Consumption (Tom's Hardware) http://www.tomshardware.com/2007/01/08/can-amds-65-nm-core-fight-back/page8.html [i]"...Generally speaking, all new 65 nm Athlon 64 X2 processors require less energy. Our test results show that the energy savings between a system with a Windsor processor and a Brisbane-equipped machine at 2.6 GHz is 2% in idle mode with Cool'n'Quiet enabled (Brisbane offers the Enhanced Halt Mode C1E), 5% in idle without Cool'n'Quiet, and as much as 11% under full load. The last number especially makes clear the huge difference that the new processor makes in energy consumption!..."[/i] [i]"...Whether you run a 90 nm Windsor core or the 65 nm Brisbane newcomer only makes a difference on your energy bill: under load, the new processor is capable of reducing the power requirements of the overall system by up to 11%. This is a tremendous improvement given that the rest of the system remains unchanged, and the performance is very much the same as well..."[/i]

TechExec2
TechExec2

. [b][i]"...I am not going to buy a worse CPU just for pitty or to preserve competition..."[/i][/b] Don't ever buy from AMD for "pity". But, when AMD offers the same or better price-performance, it is very smart to [u]always[/u] buy from AMD. There are lots of people mindlessly buying "Intel Inside". The smart move is to get the same or better price-performance from AMD while helping to keep AMD in the game [u]for free[/u]. Without AMD competing with Intel, the price of Intel processors will rise dramatically and "Intel Inside" buyers will have nobody to blame but themselves. A true definition of dumb is not thinking past today, buying Intel when AMD provides the same or better value, and paying the price tomorrow. Dumb. Really dumb. [b][i]"...If they want to be competition to Intel they have to release as good CPU's as them or better (which, at least currently, they don't)..."[/i][/b] This is completely false. See (1). ---------------------------------------- (1) ALWAYS use an AMD CPU when you build your PCs *** UPDATED *** http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=218246&start=0

georgeou
georgeou

Recently, AMD slashed the price of their X2 processors by around 40%. That means if you're not going to overclock and you don't intend to keep your CPU busy (less efficient on X2), it's a better deal to go with an Athlon X2 processor. This article was written before this transpired though. So we have a situation where AMD is willing to bleed financially and sell a low-yield low-profit 90nm 2.8 GHz X2 processor at only a slightly higher cost than an Intel 1.83 Core 2 which is very profitable and high-yield. AMD is desperate to get some volume going because not having volume means death. AMD lost over $600 in Q1 2007 and they?re just offering these prices to stop losing market share to Intel. As far as I?m concerned, AMD X2 processors are a good buy for most people. I still like the Intel 965 chipset more for the advanced features but most people will do fine with a cheap AMD X2. Note that the price differences on the chip itself isn?t necessarily enough of an offset to get people to go AMD over Intel when you?re looking at the cost of the whole system. But I agree with you 100% that buying AMD for the sole purpose of keeping an inferior product and company alive is dumb. When TechExec posted his message, the X2s hadn't droped that much in price yet and I completely disagree with his post.

georgeou
georgeou

You're right. We just take the $122 price out of it if you want to run Linux. We can shove a dirt cheap 19" display at $150 and knock the system down to $860. We can use a dirt cheap and much slower single core CPU and motherboard combo for $80 and shave $200 off the price and get it down to $660.

TechExec2
TechExec2

. [b]It's counter-intuitive[/b] Like many things, the smart choice is often counter-intuitive. The majority might be inclined to buy the Intel CPU. But, in order to preserve the pricing and competition [u]beyond just today[/u], you must buy the AMD CPU instead. So, as long as AMD is price and performance competitive, and much smaller financially than Intel, AMD is the smart choice. There will be plenty of other people who just don't get the smart counter-intuitive logic of this and Intel will be just fine. The smart people will be saving the not-so-smart ones from themselves (as is so often the case in life).

tep0583
tep0583

When I built my former PC, I did a lot of research into inexpensive mobos. The ECS K7S5A series of mobos came up as inexpensive mobos that were excellent runners. I have a K7S5A and a K7S5A Pro, both of which have given my years of trouble-free service. They weren't particularly suited to overclocking, but I was content to forgo that ability for stability. I was not disappointed,

TechExec2
TechExec2

. I've been using AMD exclusively since the K6 too! Great value and performance. Without AMD and the Athlon 64 CPU, Intel would never have been motivated to create the Core 2 CPU. Those shortsighted (blind?) Intel fanboys should never forget that.

jck
jck

My next one will be soon in the threads. I decided to spec a game machine, rather than home office to see if I could do as well as George. BTW...I read somewhere that AMD has a 40-50nm process in the works that will reduce the power consumption and thermal level even more. Anyways...off to show my PC spec abilities. cheers :)

cls
cls

If they're on a budget, maybe they care about their electric bill. We pay about 15 cents per kWh here. That runs the 200 watt PC for 5 hours. 45 cents in a house with conventional air conditioning. $13.50/month. Call it $35/month to leave it on all the time, since the A/C usually isn't running at night. The tests cited in this thread are so flawed it's hard to know where to begin. How efficient is the DC/DC converter on the motherboard? 70%? 85%? Are both test boards equally efficient? Is "rendering" completely CPU-bound, or does it represent a real-world address pattern? How much of the heat in the test machines comes from the video cards? Semiconductor processes vary quite a lot: chips made at the end of a run can be 3x faster-cooler than the same exact device from the beginning of the run while the fab is still being tuned. (Among samples of the same device from the same fab, the faster ones run cooler, because they spend a greater fraction of their lives waiting quietly for the next clock tick.) Were the test chips early production, or late? One sample is just not statistically valid. When I designed CPU boards, I had to design for the worst case of every parameter in the data sheet. Worst operating ambient temp, worst e-sub-r, slowest RAMs... Also the best case, to guarantee hold times. "Typical" just means half the parts aren't that good. The data sheet parameters were the only thing the manufacturer was willing to stand behind. The difference between "typical" and worst made my designs robust, and we still had to manufacture if we got a shipment of near-worst-case parts. I wonder if anybody still does things that way.

TechExec2
TechExec2

. [b][i]"...Although Georges example from TechReports has some apparent flaws that we all agree to..."[/i][/b] Apparently you and I agree there are some flaws in those power use tests. George has yet to agree with this. [b][i]"...Tech, do you offer any tests that can put anything you're saying as an example as to "how fast and how cool like Intels" chips are?..."[/i][/b] Yes on the performance. No on the power use. Performance: My post here (1) matches up the AMD and Intel processors using the PC WorldBench. Test results are cited. Power: My post here (2) lists the power ratings provided by both manufacturers. I do not know of any testing done by 3rd parties that agrees, or disagrees, with AMD's and Intel's claims. And, I have no reason to disbelieve either one. If one of them were way off, I would expect someone to nail them on it. So, I generally trust both of them are being truthful. [b][i]"...As far as I've heard, Intel still has the upper hand in this area of speed..."[/i][/b] Not exactly. It's true that Intel's current fastest desktop CPU (introduced last year) is faster than AMD's current fastest. But, AMD currently has CPUs that match Intel CPUs at many performance points below that. Before the Intel Core 2 was introduced last year, AMD had the fastest desktop CPU crown for a few years. Later this year, AMD CPUs based on Barcelona will likely retake the "fastest CPU" crown. Intel's plans show that Intel may retake it again in 2008. And so it goes. Intel's Core 2 Duo has better performance-per-clock than AMD's Athlon 64 X2. This means that the AMD CPU with matching performance runs at a higher clock speed. For example: AMD Athlon 64 X2 5000+ (2.60 GHz) matches Intel Core 2 Duo E6400 (2.13 GHz). [b][i]"...so if they're running cooler, they're just running slower too..."[/i][/b] Yes and no. Generally, the higher the clock frequency, the higher the power consumption, and conversely. But, this varies depending on CPU design. A CPU with more transistors will generally use more power than one with fewer transistors. For example, a CPU with a larger on-chip L2 cache will typically draw more power than one with a smaller on-chip L2 cache. And, a larger L2 cache will generally enable a CPU to perform better. -------------------------------------- (1) ALWAYS use an AMD CPU when you build your PCs *** UPDATED *** http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=218246&messageID=2209881 (2) AMD's latest CPUs run fast and cool, just like Intel http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2180650

jakesty
jakesty

Although Georges example from TechReports has some apparent flaws that we all agree to, the information you state is just that, no tests done in a 'real world' or any world setup. So just based on the fact that the info is posted really has no merit in how hot or how many watts are even used. Tech, do you offer any tests that can put anything you're saying as an example as to "how fast and how cool like Intels" chips are? Did anyone atleast put a thermometer on the CPU to get "something"? As far as I've heard, Intel still has the upper hand in this area of speed, so if they're running cooler, they're just running slower too.

TechExec2
TechExec2

. I've stated my objections to this very clearly here (1). I'm not going to repeat all of them. Readers here have to be idiots, or smoking something, in order to buy your false statements about this. [b][i]"...Those 90nm AMD chips are NEW!..."[/i][/b] Of course they are. That doesn't take anything away from my points whatsoever (1). If you want to say that the 90 nm Athlon FX-7x CPUs, and the 90 nm Athlon 64 X2 6000+ CPU, are less power efficient than the 65 nm Intel Core 2 Duo CPUs are, fine. That's the truth. I agree. But, these blanket statements from you [i]"...AMD chips consume a little more power in idle..."[/i] and [i]"...AMD chips consume a LOT more power at peak workloads..."[/i] are false. The truth is that AMD's latest 65 nm power-efficient designs compare very well with Intel's latest 65 nm power-efficient designs (2). Except for the 65 nm Athlon 64 X2 5000+, TechReport compared the [u]worst[/u] of the current AMD CPUs to compare with the best of Intel's (vis-a-vis power usage). And, you cited them to support your false blanket statements. I have some strong points. I stated them clearly (1). You're ignoring them and filling this thread with blanket statements supported by severely flawed testing. Nobody who wants to have an honest discussion does that. You should just admit you made a mistake in using TechReport's flawed testing to support your broad claims. Until you come up with some solid evidence, you shouldn't make claims like that at all. --------------------------------------------------- (1) Interesting but severely flawed article http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2220261 (2) AMD's latest CPUs run fast and cool, just like Intel http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2180650 [b]AMD "Brisbane" CPUs:[/b] AMD Athlon 64 X2 3600+, 1.9 GHz, 2x512KB L2, 65 nm, 65 watts TDP AMD Athlon 64 X2 4000+, 2.1 GHz, 2x512KB L2, 65 nm, 65 watts TDP AMD Athlon 64 X2 4400+, 2.3 GHz, 2x512KB L2, 65 nm, 65 watts TDP AMD Athlon 64 X2 4800+, 2.5 GHz, 2x512KB L2, 65 nm, 65 watts TDP AMD Athlon 64 X2 5000+, 2.6 GHz, 2x512KB L2, 65 nm, 65 watts TDP [b]AMD "Windsor" CPUs[/b] AMD Athlon 64 X2 4600+, 2.4 GHz, 2x512KB L2, 90 nm, 89 watts TDP AMD Athlon 64 X2 5000+, 2.6 GHz, 2x512KB L2, 90 nm, 89 watts TDP AMD Athlon 64 X2 5400+, 2.8 GHz, 2x512KB L2, 90 nm, 89 watts TDP AMD Athlon 64 FX-62, 2.8 GHz, 2x1MB L2, 90 nm, 125 watts TDP AMD Athlon 64 X2 5600+, 2.8 GHz, 2x1MB L2, 90 nm, 89 watts TDP AMD Athlon 64 X2 6000+, 3.0 GHz, 2x1MB L2, 90 nm, 125 watts TDP [b]Intel "Conroe" CPUs[/b] Intel Core 2 Duo E6300, 1.86 GHz, 2MB L2, 65 nm, 65 watts TDP Intel Core 2 Duo E6400, 2.13 GHz, 2MB L2, 65 nm, 65 watts TDP Intel Core 2 Duo E6600, 2.40 GHz, 4MB L2, 65 nm, 65 watts TDP Intel Core 2 Duo E6700, 2.66 GHz, 4MB L2, 65 nm, 65 watts TDP Intel Core 2 Duo Extreme X6800, 2.93 GHz, 4MB L2, 65 nm, 75 watts TDP Intel Core 2 Quad Q6600, 2.4 GHz, 8MB L2, 65 nm, 105 watts TDP (2 separate dual cores) Intel Core 2 Quad QX6700, 2.66 GHz, 8MB L2, 65 nm, 130 watts TDP (2 separate dual cores)

georgeou
georgeou

Tech, you're complaining about the usage of 90nm AMD chips? Those chips are brand new! AMD hasn't finished their 65nm transition and they won't finish for another year! AMD won't even convert their notebook Turion chips to 65nm until Q2 2007. You need to blame AMD for the existence of brand new 90nm parts.

georgeou
georgeou

"Techreports attempts to show actual usage, but I have "some" reservations about that given how they tested the system. They say they include all of these other parts which I'm sure were the same on both motherboards, but ultimately the M/B is a variable. One in how efficient it operates with the CPU and even more so in how the south/north bridges operate with the hard drives and other components." Here's the problem Jakesky. AMD would have us believe that Intel Motherboards use even more power than AMD-based Motherboards because you need a power-hungry memory controller for the Intel boards. That's because Intel CPUs don't have the memory controller built in to the chip (not till Nehalem). So if anything, the Motherboard advantage "should" go to AMD. Your concern about the use of different Motherboards actually favor AMD in theory but they're still coming out worse. The reality is that the chip dictates the Motherboard you use. The Motherboard is part of the platform.

TechExec2
TechExec2

. [b][i]"...Techs numbers look like the ones you'd get from their website..."[/i][/b] Yes. That's where they came from. I got them from AMD's and Intel's websites and I make no secret of that. Clearly, those numbers are just the manufacturer's ratings. Measuring actual CPU power usage would be most interesting. Unfortunately, [u]that is not what TechReports did[/u]! They measured [u]total system power draw[/u], and that is not the same as CPU power draw at all. [b]The biggest problem with TechReports power use tests[/b] There are many problems with TechReports power use tests. But, the biggest one is they are comparing Intel 65 nm CPUs with one AMD 65 nm CPU and several other AMD 90 nm CPUs. It is a ridiculous comparison. It is just as ridiculous as comparing the new AMD 65 nm CPUs vs. the old Intel Pentium D 90 nm CPUs or old Pentium 130 nm CPUs.

jakesty
jakesty

Techs numbers look like the ones you'd get from their website, not how they really are in operation. Techreports attempts to show actual usage, but I have "some" reservations about that given how they tested the system. They say they include all of these other parts which I'm sure were the same on both motherboards, but ultimately the M/B is a variable. One in how efficient it operates with the CPU and even more so in how the south/north bridges operate with the hard drives and other components. There could be some prejudice on one system or another. Just because it says it's 65 watts doesn't mean that you turn on the computer and BAMM, 65 watts are being used, just like TechReports show. I have also seen this kind of result from an E6600 we have that shows an entire system at idle using 71 watts, jump substantially when in use. Lastly, even if you took the numbers on the surface that the AMD cpu is 65 watts and so is the Intel, for the same highest end CPU the Intel is still a faster solution. (AMD 5200 vs E6700) And that's what it comes down to, GHZ per watt of power. How efficient a CPU is per instruction or cycle, given the same request. Right? One may use slightly more or less, but how efficient did it use the power to get there?

TechExec2
TechExec2

. [b][i]"...AMD chips consume a little more power in idle..."[/i][/b] False! [b][i]"...AMD chips consume a LOT more power at peak workloads..."[/i][/b] False! [b]Fact: AMD and Intel are comparable in power consumption[/b] The fact is, the latest 65 nm AMD and Intel CPUs are very comparable when it comes to power consumption. If you want to include AMD's older 90 nm CPUs in a power use round up, you should also include Intel's older 90 nm CPUs. I'm quite sure that most readers can see through these shameless attempts to misrepresent the facts about this. Shame! Shame! X-( [b]Utter nonsense from you[/b] What complete utter nonsense from you, George Ou. I merely cited power ratings from AMD and Intel. They are not my numbers. TechReport's power consumption tests were severely flawed and should be completely disregarded (1). ------------------------------------- (1) Interesting but severely flawed article http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2220261

TechExec2
TechExec2

. Too bad you cannot refute what I posted. As usual, you merely divert attention to something else. You must also be good at dealing Three Card Monte (1). :^0 TechReport's power consumption tests were severely flawed and should be completely disregarded. [b][i]"...The reason "EE" chips from AMD aren't tested is because they're not available to the open market; only to a few OEMs. Those "EE" chips are lower-clock and they're basically slower than the Intel Core 2 Duo laptop processors. So if you want to compare the EE chips from AMD, compare them to Intel's mobile line..."[/i][/b] Irrelevant! I made no reference to this. Either cite test results or this diversion of yours is completely irrelevant and should be disregarded, just like the TechReport power tests. ---------------------------------------- (1) Three-card Monte http://en.wikipedia.org/wiki/Three-card_Monte

georgeou
georgeou

Too bad AMD won't even make your claims Tech. I asked them about these results and they changed the subject to servers. Those tests are legit. The reason "EE" chips from AMD aren't tested is because they're not available to the open market; only to a few OEMs. Those "EE" chips are lower-clock and they're basically slower than the Intel Core 2 Duo laptop processors. So if you want to compare the EE chips from AMD, compare them to Intel's mobile line.

TechExec2
TechExec2

. [b][i]"...The simulated processors' voltage may not be exactly the same as what you'd find on many retail E6600s and E6700s..."[/i][/b] How can you run a power consumption test when you're not even using the same voltages as the real CPUs? This admission by the author tells you all you need to know. These power use test results are INVALID. [b]Severely flawed power consumption testing[/b] :0 I like what the author tried to do in this article (2). But, his power consumption testing is severely flawed because: 1. He did not use the latest low-power AMD CPUs in his tests. Everyone admits you can heat your house during winter with the Athlon FX-74. 2. He "simulated" the Intel CPUs that he did not have. The author admits he did not use the same voltages as the real CPUs. Further, these Intel CPUs are not all exactly the same! A larger L2 cache draws more power. Merely adjusting CPU clock speed does not make for a valid "simulation". 3. His tests measure [u]system[/u] power use, not CPU power use. How do the motherboards compare? Chip sets compare? Memory compare? Power supply efficiencies compare? Video cards compare? Hard drives compare? As written, the author's power use testing is [u]severely flawed and should be completely disregarded[/u]. The only value of this article's power use testing is to motivate a better test. But, I really like what he tried to do. [b]Closest power use comparison from this test[/b] The closest thing to a reasonable power use comparison in this article is AMD Athlon 64 X2 5000+ 65nm vs. Intel Core 2 Duo E6400. These two CPUs perform comparably (1). In the author's test, at idle, the AMD CPU drew 1.6% more power. At load the AMD CPU drew 31% more power. But... The author's test results are highly questionable. Why does the E6400 draw 9.3% *LESS* power than the E6300? This makes no sense and must be caused by the problems in his testing that I cite above. If you reposition the E6400 between the E6300 and E6600 where you would expect it to be, the E6400 would draw about 158 at full load. This makes the comparable AMD CPU draw only 18.3% more power at full load. But, once again. I throw the entire thing out. Testing the actual Intel CPUs could make things better for AMD, or worse. We just don't know. [b]Peak power use is interesting but...[/b] Peak power use is interesting. But, a desktop processor typically spends most of its time at idle. [u]It would be pretty dumb to buy a desktop CPU based on peak power use[/u]. ----------------------------------------- (1) ALWAYS use an AMD CPU when you build your PCs *** UPDATED *** http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=218246&start=0 (2) Power consumption and efficiency http://techreport.com/reviews/2007q1/cpus/index.x?pg=13

TechExec2
TechExec2

. The newest AMD CPUs use less power, just like the newest Intel ones (1). AMD offers better price-performance than Intel (2). In particular, the fastest AMD desktop processor ever made (Athlon 64 X2 6000+) can be had for only $239 (3). This processor is so fast that it was not until July 2006 that Intel had a faster one (Core 2 Extreme X6800). Right now is an incredible time to build a system with an AMD processor. P.S. If you need more speed than the Athlon 64 6000+ offers, you can choose an Intel CPU like the Core 2 Extreme X6800 ($965) or Core 2 Extreme QX6700 Quad ($967) (4). ----------------------------------- (1) AMD's latest CPUs run fast and cool, just like Intel http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=212361&messageID=2180650 (2) ALWAYS use an AMD CPU when you build your PCs *** UPDATED *** http://techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=218246&start=0 (3) AMD Athlon 64 X2 Dual-Core Processor 6000+ $239 http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=80743 (4) Intel Core 2 Duo CPUs http://www.zipzoomfly.com/jsp/ProductList.jsp?ThirdCategoryCode=110514&SortBy=D

TechExec2
TechExec2

. Thanks for your post. You are one of the smart ones!

stew
stew

Your point is excellent, but it probably doesn't apply to the audience of this discussion, does it? That is, those looking to build a great PC on a budget probably don't have a data center.

cls
cls

If AMD's dual-core and Intel's dual-core are within 15% of each other in speed, it's time to look at other variables. In the data center, you pay for electricity three times: once in your server and twice in your air conditioner. AMD's been advertising dual-cores under 65 watts. How much heat does Intel's dual-core produce doing the same job?

stew
stew

Intel never innovated so well as when AMD stole their thunder. AMD is behind Intel now, which they hadn't been for several years. Intel made a better marketing move, which has bought them time and money to produce a better quad-core CPU, while providing some improvement in the present. Still, I think it is wise to prefer AMD over Intel because so many remain uncomfortable with anything-but-Intel (thanks to "Intel Inside," etc.). If you *need* the fastest system possible for the money and Intel is the fastest, buy Intel. If you don't *need* the fastest and AMD is fast enough for the intended use of the system, then choose Intel only if they are cheaper and faster or significantly cheaper for the same performance. Obviously, if money is your main concern, then you have to find a CPU that is fast enough at the lowest price you can get, whether that favors Intel or AMD. If AMD fails, we all lose, so a little deference to them is good. It will help to keep the game alive.

PhilippeV
PhilippeV

Why would you need to constantly update Windows on a PC that is used for playing games (I mean here, local games, not online games) or a local media center? My suggestion was about reconverting an old PC for such other local utility. If it works well locally, it should not suffer from security risks. You don't need to update constantly a PC that is not on the first front of Internet, provided that you don't let it connect to external sites. And no, Linux does not run most games... It will be great for surfing, for your free Star Office suite, or for your local firewall and for receinving/sending emails. But anyway you still can't do everything today. I don't say that Linux is bad, in fact it is even better and ssafer for many apps, but not all. It's not unreasonable now to have multiple systems, and put the most exposed system out of the battle frontline for Internet security risks. And if you plan to manage your personal bank accounts, better keep your files on a PC that you'll NEVER connect to the Internet (you may use a USB key or CD-RW (or even a floppy if you still have a drive for it) to transfer some data files collected from another PC, but beware of possible infection of your USB key with autorun files). I am as much opposed to the "All Windows" rule, as to the "All Linux" one. Think differently, and use the system that just performs the expected functions, and don't install unnecessary addons. Big Pcs for making all is now much less needed. Small is beautiful, because each one is easily manageable. And there are now cute devices to perform data exchange between PCs tht do not require letting them always on or always siltently connetd with permanent shares between each other. Isolate each risk. Note that Linux also requires security updates... Don't think that spammers and criminals don't know how to use them, they do use them and are attacking bank servers or websites with some success, despite they have professional teams monitoring them. You alone at home, are probably less equipped to monitor all what is happening in your PC, even if it is less exposed to security risks than Windows). Remember that it's sufficient for spammers to succeed ONCE. Most of their strategies will be failing, but sooner or later, with the so many tricks invented each day, they will be able to introduce something in your system. The question to ask is: which data is exposed, and what is its value? Do I have backup solutions? Can I lower the risks or the consequences of damages? Can I get some protections or some help after a disaster? Can I recover from damages easily? If you have placed everything on the same system, your dependeny about it is high, and the risk to which you are exposed are also high. If you have multiple PCs, you can greatly reduce the risk by using the right data on the right place, and by multiple defence lines. i absolutely don't know anyone that has never been exposed to privacy data theft (just look at how emails are used, and linked with your online web activities, then consider a Google search about yourself! It's interesting to know what others can know easily about you, and how they can create social engineering to defeat your supposed security, unless you know how you or your contact can ensure that you are effectively the one that you pretend to be, or easily detect harvesters. I have in fat not favored any OS in my comments you criticize here. I gave some hints about reusing the tools you have and calculating the cost, and measuring the efficiency of the solutions you have. These were just hints, not general recommandations.

jackie40d
jackie40d

Why would you keep on jumping into the Fire and loading a O/S that needs its tires fixed EVERY MONTH or a CARB replaced EVERY MONTH . . Would you buy a car that way ? WELL ? grow up and get a computer put it together and add one of the Linux versions add Code Weavers. . And you can run 90 % of your windows programs . .

PhilippeV
PhilippeV

Microsoft knows that a lot of games won't run on Vista, beause of its current lack of support of OpenGL (this will come later, when AMD/ATI and nVidia/GeForce will upgrade their graphics drivers); many gme programmers have no intention to port their games now to DirectX 9, given that their games work faster with openGL and are easier to port on other platforms (Mac and consoles) with OpenGL, so they will continue to develop first for consoles, then for PCs, by starting with OpenGL (which is also supported on their very expensive graphics workstations where they model and develop the graphics, and that only support OpenGL, the true industry standard). Tht's a good reason why Microsoft is still shipping XP which gives infinitely better performance for gaming (3D and video). The small reduction of performance in other domains is not important and compensated by the always growing performance of processors; Vista will remain a good Windows because it will handle more memory and will run more securely (it is the new memory isolation model of Vista which causes the small reduction of performance, but this is for increased security everywhere in the OS). ATI and nVidia have both promissed the support of OpenGL on Vista (it's not Microsoft that will provide it, but Vista is ready to accept OpenGL with an API compatible with the existing one in XP). What Vista will provide however, that is not in XP, is its new management for multiple virtual frame buffers, and a fast coloring engine with transparencies between multiple applications, something that OpenGL does not provide natively exept in a single window. Vista will allow hardware implementations of OpenGL to work in virtual frema buffers, and will allow cohabitation of multiple transparent windows mixing different rendering technologies (DirectX including the legacy GDI, or OpenGL) and fast virtualization of graphics devices (so no more exclusive mode where the mouse cursor and keyboard is grabbed and forbids access to the display by the OS). Games will become playable in multiple consoles, for example through remote connections from another PC on the LAN or on Internet. As long as there will be no offer that supports OpenGL on Vista, XP will remain at its current high price. After that, the upgrade to Vista will be worth the value, and XP will start decreasing in usage and in pricing at Microsoft. XP will then be a great OS for convering older PCs into file servers, media servers for your salon or HiFi system... or as a bakup unless you choose to convert it to Linux only as a utility server which hosts your Internet firewall and just serves your medias at home (note however that DRM-protected medias will not play correctly on Linux, unless it is tweaked or you choose a expensive commercial distrib that includes the necessary licences and a subscribed service; an old version of XP will be useful in the future, as a Media center for DRM-protected files, but definitely not to run a firewall or a backup, so it will only be a solution for a third PC or as a game console...)

georgeou
georgeou

Second, don't need a 450 watt PSU when this system only uses 110 watts peak.

jackie40d
jackie40d

If you buy the big windows its like $400.00 and Xandros was only $90.00 deilvered to my door ! ! Thats $310.00 less right there And I get Motherboards with at least 4 slots for memory and can take faster CPU's of same kind so they can up grade and I put in a middle of the line CPU close to 2.7 gigahertz close to $100.00 give or take $20.00 depending on the person and their wallet I like to have them add at least 1 gig of ram so they are not dragging hind leg . . Then add what they want for Hard drive in a case with a 450 watt Power supply ! Then add burner and graphics card keyboard and mouse . . Let them go shopping for monitor ( so they see what it costs ) I try to talk them into a version of Linux since it does not have PATCH TUESDAY to worry about and Linux has a much better anti virus and fire wall . . And hackers do not mess in their own back yard . . not cool to do hahaha !

jakesty
jakesty

But, I believe Georges pretense was building a Vista capable system, not something that grandma will only use for email.

jackie40d
jackie40d

Well I will check on it and maybe change brands I did a copy and paste to a file to keep for refference . .

TechExec2
TechExec2

. [b]It is confusing[/b] Years ago, AMD and Intel used a similar architecture and a higher clock speed meant a faster processor, even when comparing AMD vs. Intel. Those days are gone for both AMD and Intel. The CPU clock speed is much like the RPMs of the engine in your car. Those engine RPMs do not determine the speed of the car. The transmission plays a big role in that. Both AMD and Intel have been making design changes (to the "transmission") that increase the performance per clock. So, a lower clock speed can produce higher performance, just like a lower engine RPM in 4th gear can make the car run faster than a higher engine RPM in 1st gear. Example: The current Intel Core 2 Duo E6300 (runs at 1.86 GHz) performs the same as an AMD Athlon 64 X2 4600+ (runs at 2.4 GHz), and they both perform like a Pentium running at 4.6 GHz. [b]Use Benchmarks[/b] Clock speed means nothing now except for when comparing CPUs in the same line from the same manufacturer. So, I prefer to use the PC WorldBench benchmark test to evaluate the speed of the CPU. Look for reviews on the web, they're easy to find on Google. This benchmark uses real applications and measures the performance that you will actually see. Other benchmarks measure the number of "gigaflooops" and "megabeeeps". Well, I cannot see those, can you? :-) [b]AMD's Model Numbering Scheme Tries to Help[/b] AMD's model numbering scheme is designed to make it easy to compare the relative performance of the AMD CPUs vs. Intel [u]Pentium[/u] CPUs. It works like this: AMD 3000+ performs like a Pentium 3.0 GHz CPU AMD 4000+ performs like a Pentium 4.0 GHz CPU AMD 5000+ performs like a Pentium 5.0 GHz CPU AMD 6000+ performs like a Pentium 6.0 GHz CPU ...etc... [b]In conclusion[/b] Now that you know the "secret", it's easy to see that AMD's model numbers provide the [u]easiest[/u] way to know the performance you're going to get. If you buy an AMD Athlon 64 X2 6000+, it will perform about twice as fast as your current Pentium 4 (3 GHz). You have to dig to find out which of the new Intel Core 2 Duo CPUs will do that. All the best to you!

jackie40d
jackie40d

When I buy a Intel 3.0 gigahertz P4 I get a CPU which is doing 3.0 gigahertz then when I buy a AMD ( their numbers do not make sense ) I have to read spec's to see what its really doing like a 2400 something is only doing 1.6 gigahertz I like to know what its doing when I buy it, not have to read all the fine print !