What AMD Processor will convert faster

[txpharoah]

Ladies, I think that you are all getting a little bit heated! This reminds me of arguments I used to have in the school playground about whether He-Man could beat The Incredible Hulk in a fight. ..... And to settle the argument once and for all, the Incredible Hulk would kick He-Man’s ass!

A bit after my time, but I'd like to see that fight. A Wrestlemania special match maybe. The NEW Cartoon Network Heman vs the NEW movie Hulk. Very nice.

Yeah pablo, I just share experiences. It's up to everybody else on how they decide to use the information I give out. I work in the media business, and that's how it goes: some care, some don't, some listen, some don't. I still get paid the same. Or in the case of these forums, still DON'T get paid the same.

[WeedVender]

On CNN they had an interview from one of the CG programers in Starwars attack of the clones and he stated that they use nothing but AMD Atholon MP processors for the movie....

[Metaluna]

Well I have both an AMD and a P4 system sitting here next to me, and I've been meaning to benchmark them, so I figured this would be a great time. It isn't the most scientific, controlled benchmark in the world but I think it's pretty reasonable.

I started by capturing a 90 second video clip on my Canopus ADVC-100, saving it as a type 2 AVI. The systems under test are configured as follows:

P4 System:
2.53GHz Northwood w/533MHz frontside bus, 166MHz DDR memory bus
Asus P4PE motherboard (Intel 845PE Chipset)
1 GB PC2700 DRAM (2 DIMMs)
Windows XP Pro SP1a
Geforce 4200ti video card

AMD System:
Tbred B 1700+ overclocked to 1.83 GHz, 166MHz DDR frontside and memory bus
Epox 8RDA+ motherboard (Nvidia NForce 2 chipset)
1 GB PC2700 DRAM (2 DIMMs in dual-channel configuration)
Windows 2000 Pro SP3
ATI Radeon 8500 video


Both systems are running
TMPGEnc 2.512.52.161. For the encoding, I used my usual settings, which is 2-pass VBR, 5600 average, 8000 max bitrate, 10-bit DC precision, high quality motion search, no filters, no audio compression, elementary stream output. I also turned off the video preview mode in order to keep the video cards out of the picture as much as possible. I verified that both processors were correctly detected and had their special instruction sets turned on (i.e. MMX, SSE, and 3DNow! for the AMD, and MMX, SSE, SSE2 for the P4)

The input AVI file is DV encoded at 720x480 and is 344,702,976 bytes long. I have the Mainconcept 2.1 demo DV codec on both systems in case that matters. The resulting .m2v file is 63,114,378 bytes long on both systems. I used an MD5 checksum program to verify that the two test platforms produced identical files.

Here were the results in minuteseconds

2.53GHz P4 Northwood: 7:15
1.83GHz AMD Tbred B : 6:45

The Tbred is actually about 7% faster than the P4, despite the P4's significant clock speed advantage and the presence of SSE2. I was really surprised. I was expecting the P4 to be much faster. I suppose Windows XP could be slowing things down a bit, and the P4 is my primary PC, so it has more stuff installed on it, but I checked the Task Manager and TMPG and the System Idle process were the only two major users of CPU time. I'll go in and check the BIOS settings on the P4 after this but I'm pretty sure I have all the memory settings set to "turbo" and the memory bus at 166/333 MHz.

Now obviously I'm comparing an overclocked system to a non-overclocked one, but at 1.83GHz the Tbred B is running at the speed of a Barton 2500+. The Barton should be at least as fast, and probably a little faster. Note that the P4 2.53B is running about $175 at Newegg.com while the Barton 2500+ retail is selling for $96. The Epox 8RDA+ motherboard is about $85 and the Asus P4PE/L is $104. All other component prices should be equal. Thsi means the P4 solution was $98 more than the AMD. Not a huge amount compared to overall system cost, but you could buy a second Barton as "burn up insurance" and still have a potentially faster system than the P4. In my case I only paid $50 for the AMD CPU and around $229 (at the time) for the P4 so the price difference is much greater.


As for AMD's burning up....I've assembled and used dozens of them (including 3 duals that we run 24/7 at work, one of which is a prototyping testbed that we use to debug our PCI card designs). I admit I'm not a highly paid tech (just a moderately well paid electrical engineer), but I have seen exactly one AMD processor burn up, and that was a Tbird that I had been messing around with. I managed to tear off one of the foam rubber stabilizer pads, and I tried to reattach it with superglue. Well the foam rubber sucked up the glue and became hard as a rock, preventing the heatsink from seating properly. Really, AMD cpu burn is just not that much of a problem. In fact the only time you ever hear anyone bring it up is when an Intel advocate talks about it. Most AMD people scratch their heads and wonder what they're talking about. It just isn't a big problem. Go to any AMD-centric website and look in the user forums if you don't believe me (I suggest AMDMB.com). It just doesn't come up that often. I would be much more concerned about power supply failure, which can take your CPU, motherboard, and a whole lot more with it when it dies. Replacing a $95 cpu shouldn't be that much of a hassle compared to a motherboard swap, but again I've never seen it happen on a working system so I can't say.


If you're really worried about your heatsink falling off (again, have never seen it happen on a working system...have seen them come off on PC's that had been shipped UPS though), just get one of those heatsinks that bolts onto holes in the motherboard instead of clipping to the socket (e.g. Thermalright SLK800U). I have one on my P4 and that thing isn't going to go *anywhere* without the liberal use of a sledgehammer.
[/list]

[txpharoah]

The Tbred is actually about 7% faster than the P4, despite the P4's significant clock speed advantage and the presence of SSE2. I was really surprised. I was expecting the P4 to be much faster. I suppose Windows XP could be slowing things down a bit, and the P4 is my primary PC, so it has more stuff installed on it, but I checked the Task Manager and TMPG and the System Idle process were the only two major users of CPU time. I'll go in and check the BIOS settings on the P4 after this but I'm pretty sure I have all the memory settings set to "turbo" and the memory bus at 166/333 MHz.

Be sure the TMPGenc setup has SSE2 and all enabled. They often come disabled. Be sure the system drags are the same (network, anti-virus, etc). And make sure they have the same OS for the tests. It's good to tests, but the configurations must match exactly. Are the same codecss being used. Etc. Sounds like you did some of this already, but something seems off, can't quite place my finger on it. Oh well. The encodes seem decently fast, so enjoy both systems. Just keep an eye on the AMD if it has unattended extended-use-sessions.

[txpharoah]

On CNN they had an interview from one of the CG programers in Starwars attack of the clones and he stated that they use nothing but AMD Atholon MP processors for the movie....

That's because AMD processors perform better with CG (this is not video). And they use Maya, which is apparently 3Dnow enhanced. Got a friend that use that at his job. I've gotten my ear-full of the Star Wars EpI/II in past years from this guy. But they also use SGI and other systems on a network that links processors and power, and all kinds of stuff that makes no sense to a non-tech like myself, no matter how many times he explains the process. They are not home system setups, that's safe to say.

[alucard2050]

my 2 cents.....

i have used both pentium and amd chips...i have stuck with amd for almost 3.5 years now...i will never use a pentium chip again...i have never had a chip melt on me....pentium used to give me CONSTANT problems...i have a friend who works in a computer shop and he won't touch a pc with an amd chip...but, thats my personal preferance....

[LanEvo7]

From what I understand, AMD processors produce much more heat compared to Intel processors...........therefore you have blown out chips after extended use.

Your information is outdated. Newer AMD (Thoroughbred A and up) produce heat about equal to their Intel counterparts.
http://www.anandtech.com/cpu/showdoc.html?i=1718&p=4
Here is a heat comparison.

Edplayer and LanEvo7, give up the vendetta you are jumping on everyone that criticises AMD’s or contradicts anything you say.

You should re-read my posts carefully. If you noticed my first post, I told outright that an Intel CPU would perform better.

I jump on TxPharoah because he OUTRIGHT LIED to people. He isn't contradicting me, he is MISREPRESENTING the facts. He even admitted he that he doesn't know what he is talking about. Its ok for you to share experiences that you've had, but posting FALSE claims is another thing. I mean if you know nothing about automobiles, don't go around telling people what type of shocks they need. Same theory applies here. [/b]

[gmatov]

The PGA (Pin Grid Array) Socket A AMD Athlon and AMD Duron processors have very different thermal specifications than any preceding AMD processor. Due to these differences, the AMD Athlon and AMD Duron processors should NEVER be run without a heatsink, not even for a few seconds. Doing so will cause the processor to overheat and fail immediately, resulting in permanent damage. Never boot an AMD Athlon or AMD Duron processor without a heatsink and fan installed.
A thermal interface material, such as phase-change material, must always be applied between the processor and heatsink as well. Typically, AMD recommended heatsinks will include the thermal compound.

The above quotes are as of half an hour ago, from AMD Tech pages.

"If it is true that you've had that many chips melted on you, you probably didn't install it correctly. You really shouldn't blame the Chip manufacturer for your incompetence."
This quote is from Lanevo, posted June 8, 20:14, that's PM, last night. I believe that settles that I was talking about a post in which it was stated "you didn't install it correctly", no?

Weedvendor,
No, No, NO! It is NOT the CPU with a 1 degree per second temperature rise, it is the limit of ability of the thermistor to read more than 1 degree per second OF temperature rise. There is a world of difference there. Read the above. AMD says to operate for even a few seconds without proper cooling WILL cause failure, do not attempt to boot without heatsink and fan applied, as it WILL cause catastrophic failure.

Ed,
It did not say improperly installed heatsink, it said improperly installed CPU. Go back to last night's posts. Just click on the numeral 1 below. LanEvo, post time, 20:14. And, on NO page of the AMD site does it say the use of thermal compound will void the warranty. They Do recommend 6 or 7 TIMs, but in the above verbiage do tell you to go sparingly with thermal compounds, as too much will actually act as an insulator.
TechTV at one time demo'd an AMD board without cooling attached. It has been a while, but I believe it was less than 5 seconds after power up, clouds of smoke, totally destroyed CPU.
If you want a magazine less biased than the mainstream computer pubs, try Maximum PC (Minimum BS ). They will and do tell you something is a piece of crap, if it is, even if the facing page might have an ad for the product being blasted. They do see-saw on the AMD-Intel issue, as one will come out with its latest and greatest, they will test, report that it is the new leader, then the next issue, the opposition will unveil new, test, unseat the prior trophy holder.
Oh, hell. That's enough for tonight.

[/u]

[soundforbjt]

Well I have both an AMD and a P4 system sitting here next to me, and I've been meaning to benchmark them, so I figured this would be a great time. It isn't the most scientific, controlled benchmark in the world but I think it's pretty reasonable.

I started by capturing a 90 second video clip on my Canopus ADVC-100, saving it as a type 2 AVI. The systems under test are configured as follows:

P4 System:
2.53GHz Northwood w/533MHz frontside bus, 166MHz DDR memory bus
Asus P4PE motherboard (Intel 845PE Chipset)
1 GB PC2700 DRAM (2 DIMMs)
Windows XP Pro SP1a
Geforce 4200ti video card

AMD System:
Tbred B 1700+ overclocked to 1.83 GHz, 166MHz DDR frontside and memory bus
Epox 8RDA+ motherboard (Nvidia NForce 2 chipset)
1 GB PC2700 DRAM (2 DIMMs in dual-channel configuration)
Windows 2000 Pro SP3
ATI Radeon 8500 video


Both systems are running
TMPGEnc 2.512.52.161. For the encoding, I used my usual settings, which is 2-pass VBR, 5600 average, 8000 max bitrate, 10-bit DC precision, high quality motion search, no filters, no audio compression, elementary stream output. I also turned off the video preview mode in order to keep the video cards out of the picture as much as possible. I verified that both processors were correctly detected and had their special instruction sets turned on (i.e. MMX, SSE, and 3DNow! for the AMD, and MMX, SSE, SSE2 for the P4)

The input AVI file is DV encoded at 720x480 and is 344,702,976 bytes long. I have the Mainconcept 2.1 demo DV codec on both systems in case that matters. The resulting .m2v file is 63,114,378 bytes long on both systems. I used an MD5 checksum program to verify that the two test platforms produced identical files.

Here were the results in minuteseconds

2.53GHz P4 Northwood: 7:15
1.83GHz AMD Tbred B : 6:45

The Tbred is actually about 7% faster than the P4, despite the P4's significant clock speed advantage and the presence of SSE2. I was really surprised. I was expecting the P4 to be much faster. I suppose Windows XP could be slowing things down a bit, and the P4 is my primary PC, so it has more stuff installed on it, but I checked the Task Manager and TMPG and the System Idle process were the only two major users of CPU time. I'll go in and check the BIOS settings on the P4 after this but I'm pretty sure I have all the memory settings set to "turbo" and the memory bus at 166/333 MHz.

Now obviously I'm comparing an overclocked system to a non-overclocked one, but at 1.83GHz the Tbred B is running at the speed of a Barton 2500+. The Barton should be at least as fast, and probably a little faster. Note that the P4 2.53B is running about $175 at Newegg.com while the Barton 2500+ retail is selling for $96. The Epox 8RDA+ motherboard is about $85 and the Asus P4PE/L is $104. All other component prices should be equal. Thsi means the P4 solution was $98 more than the AMD. Not a huge amount compared to overall system cost, but you could buy a second Barton as "burn up insurance" and still have a potentially faster system than the P4. In my case I only paid $50 for the AMD CPU and around $229 (at the time) for the P4 so the price difference is much greater.


As for AMD's burning up....I've assembled and used dozens of them (including 3 duals that we run 24/7 at work, one of which is a prototyping testbed that we use to debug our PCI card designs). I admit I'm not a highly paid tech (just a moderately well paid electrical engineer), but I have seen exactly one AMD processor burn up, and that was a Tbird that I had been messing around with. I managed to tear off one of the foam rubber stabilizer pads, and I tried to reattach it with superglue. Well the foam rubber sucked up the glue and became hard as a rock, preventing the heatsink from seating properly. Really, AMD cpu burn is just not that much of a problem. In fact the only time you ever hear anyone bring it up is when an Intel advocate talks about it. Most AMD people scratch their heads and wonder what they're talking about. It just isn't a big problem. Go to any AMD-centric website and look in the user forums if you don't believe me (I suggest AMDMB.com). It just doesn't come up that often. I would be much more concerned about power supply failure, which can take your CPU, motherboard, and a whole lot more with it when it dies. Replacing a $95 cpu shouldn't be that much of a hassle compared to a motherboard swap, but again I've never seen it happen on a working system so I can't say.


If you're really worried about your heatsink falling off (again, have never seen it happen on a working system...have seen them come off on PC's that had been shipped UPS though), just get one of those heatsinks that bolts onto holes in the motherboard instead of clipping to the socket (e.g. Thermalright SLK800U). I have one on my P4 and that thing isn't going to go *anywhere* without the liberal use of a sledgehammer.
[/list]

Your test is not as fair a test as mine was. I used the same rated speed processor's....2.4 P4 vs XP 2400+ neither overclocked. Same exact sticks of ram....Same OS WinXP Pro SP1
Epox 8rda N-Force vs Asus P4PE not really fair (but I can't afford an 875 chipset mobo yet)

Same hard drive loaded with same apps. Same full length movie.

The Winner P4!! especially in compiling in SpruceUp!!!

[Nelson37]

Folks, my 3 co-workers and I have over 60 years combined experience fixing PC's. We've used the same 2 parts suppliers for over 10 years. There are some factors you are overlooking.

Is the failure rate, or rate of returns, (which is not necessarily the same thing, but close) for AMD's higher, both currently and over the last several years? YES, absolutely. HOWEVER, the rate declines significantly for boxed processors with matching fans, and yet more when combined with quality motherboards instead of el cheapos, and even more with pre-installed mobos and processors. AMD return rate is still higher, but much closer to Intel. I would derive from this more that AMD are difficult to install correctly rather than of significantly lesser quality. Before they added the little rubber standoffs it was quite easy to crack the chip installing the fan, for instance.

INTEL makes a significant percentage of the mobos that theie chips are installed on, versus none for AMD. Mobos using AMD chips are often geared to be less expensive, this is AMD's target market. While less expensive does not ALWAYS mean lesser quality, it very often does. There are both hi and lo quality boards for both chips, true, but more cheap boards for AMD. Also consider that AMD related products are usually designed, tested, and produced on a shorter time line in order to catch up to Intel.

I have seen scorched processors, and cracked and even explosively cracked ones, but never a melted one. What EXACTLY does "melted" mean to most people? Some of the latest AMD chips are alleviating the problem, but most of their production for the last few years runs significantly hotter than Intel. Chip fan failures are fairly common, and for another factor, yet more common in machines run well past their normal 3 - 5 year life cycle, as cost conscious buyers are more likely to do. Whether AMD's mobo-dependent heat protection becomes standard on all boards, or they perhaps change their system, is an important consideration. But now I have to check that the board supports it rather than knowing it is built into the chip.

I have used AMD since my second PC, a 286-16. I still like them for personal use. I have 3, two of them running days or weeks at a time, one of these 3 or 4 years old now. I also live in the lightning strike capital of the world, Southwest Florida. No problems, good dollar value. I usually buy 3 or 4 notches below top of the line, with a lot of salvage parts.

But for our business customers, where the cost of setup and installation, training, etc. is significant, and who are buying top-flight machines, or close, we are using mostly Intel to minimize the hassle factor.

Does anyone else see how totally invalid the above speed comparison was? Different OS, for crying out loud! Not to mention video card and almost certainly completely different installed software, other factors. Sure, you have a comparison there, but of what? Others have done more valid speed comparisons, but the latest is always a few bios revisions ago, or a new board design, or a competitive price reduction. Also reviewers are mostly gaga (sometimes for financial incentives) over whiz-bang tech most of us will never own, or which won't last 6 months. When was the last time you saw head to head speed tests with what most of us own, last year's (or the year before's) latest and greatest? No, a 2 year old review is not as good, I want the latest drivers and BIOS revisions for a real comparison - this often benefits AMD more than Intel.

Wow, way long here, no more caffeine.

[Metaluna]

Just a quick update on my test: As luck would have it I had a Win2K Ghost backup image for the P4 system, from a fresh install that I had started but never finished (long story). It is pretty much a bare install, with service pack 3 and the latest WHQL-certified drivers from Nvidia. No Antivirus software or anything like that is installed. I also double checked the BIOS to make sure all the memory and CPU timings were optimal. Everything looked good. I even pulled all non-essential PCI cards out of the system, including sound, network, and 1394 cards, and disabled the onboard audio and Ethernet as well.

First, I ran the test with no changes to the methodology. I got 7:15 again...exactly the same as under Win XP, down to the second.

Speculating that the HD might be a bit fragmented, I copied the test files to the C: partition, which, having just been freshly Ghosted, was completely defragged, and should also be a bit faster since it's using the outer tracks of the drive.

I verified that MMX, MMX2, SSE, SSE2 were all selected in the "Environmental Settings->CPU" panel. Furthermore, I went to the Task Priority menu, and set the task priorities to high, just in case some wayward service was stealing CPU cycles.

P4 Result: 7:10

Then I noticed there was a "multi-thread" option in the CPU panel that was unchecked, so I tried that too.

P4 Result w/ multi-thread + high priority: 6:52

Okay, that's pretty decent.

I tried it with the AMD and got:
Tbred w/ multi-thread + high priority: 6:21

(btw, I never bothered to disable the antivirus or defrag the HD on this system)

So, now I have the same OS (freshly installed without too much crud in the registry, etc.), no PCI devices installed, no antivirus or other CPU stealing software installed. The only differences beyond the unavoidable CPU and motherboard are the hard drive and the video card. Both machines have high-speed 7200RPM drives with 8MB cache. Intel's ATA controllers are good (IMHO), so I'm pretty confident that the P4 is up to snuff in this area. The video card I suppose could make a difference, but realistically how much is it going to affect a CPU-bound process? A few percentage points maybe?

Incidentally I don't think my results are that far out of line. Things seem to be tracking the same way on both systems based on enabling and disabling certain features, which suggests that there isn't some bottleneck that is throwing the P4 off. If the P4 results were invariant I'd be more concerned.

If you look at this benchmark (the first relevant one I could find from Google) you will see a 1.67GHz Athlon going against a 2.6GHz Northwood using TMPGEnc 2.02. The Northwood is only about 12% faster, and the clock speed gap is significantly wider than in my comparison. It could be that TMPGEnc just isn't using the SSE2 instruction set to it's fullest, or I've hit some magic set of parameters that the P4 happens to be particularly bad at or something.

http://www.vr-zone.com/reviews/Intel/P4-Northwood/page7.htm

I think that's enough benchmarking for one night. Who cares anyway. You fire off the encode and check it in the morning and see if it's done. If not, you check it again after work. Ten or twenty percent isn't going to matter much unless you're doing short clips and sitting there waiting for them.

[Don2050]

My experience......


I would like to add that I have been using AMD chips for over 2 years now doing everything from a/v encoding to gaming and never once had one go bad. Certain models of AMD run hot, ie the 1.2/1.4 Mhz T-birds, of which I had a 1.4 T-bird and though it did run hot it never failed.

I have left both my AMD comp's running for days straight doing batch encoding and never had any problems. I dont know where or why some people are saying they are unrealiable.

I have a freind who was an Intel fanboy and presented the very same arguments that AMD were unrealiable....he eventually bought an AMD after seeing that the unrealiable issue was false. I think in the past AMD did have relaibility issues..but in the last 2+ years I have owned them they have run flawlessly.


As for performance, I have heard that P4's handle video encoding better do to there pipeline architechture...was longer and let it execute non-random tasks more efficiently or something, and AMD's did gaming and office apps faster. That being said the percentages that either chip had over the other were minimal...I think in the area of 5-8% which unless you do this for a living, I cannot see the common person really noticing the difference.

anyway...just wanted to abate the AMD bashing that was going on...both have matured into good chips, performining close to equal. I think the main thing AMD has going for it is price.

P.S. Here is link to some articles on CPU architecture from Arstechnica.com if any of you are familiar. They usually do thourough reveiws on things of that sort

http://arstechnica.com/cpu/index.html

And for the original guy who asked about CPU cache
http://arstechnica.com/paedia/c/caching/caching-1.html

[q1aqza]

Sigh - Intel v AMD wars....again

A great solution would be if someone invented a dual CPU motherboard that would take one AMD and one Intel with a CPU boot option that allows you to choose which ever OS/CPU suits your needs !!!

he he I'd better get down to the patent office right away

[pablo180]

All this talk about AMD’s overheating got me worried; I have a Athlon XP 2000+ not over clocked or anything just as it came. I went into the BIOS last night to see if I have a setting that will shut down the CPU if it gets too hot (someone on here said to set it to 60c or 70c), I haven’t got anything like that on my motherboard. I checked the temperature of my CPU just to see. It had been on all day and it had been a fairly hot day but it was 12am when I checked it. It hadn’t done any encoding either that day or the night before and was only switched on that morning.

CPU: 57c
System: 27c
Fan Speed: 2500

Now that seems very hot considering it hadn’t been doing anything. I haven’t got a clue about what temperature it should be though, is that high? Or have I just got a bit paranoid with all this talk of overheating?

[IJM]

All this talk about AMD's overheating got me worried;
...
CPU: 57c
System: 27c
Fan Speed: 2500

Now that seems very hot considering it hadn't been doing anything. I haven't got a clue about what temperature it should be though, is that high? Or have I just got a bit paranoid with all this talk of overheating?

I've had my XP2100+ for nearly a year now. I have it running at 100% load for 24/7 doing encodes or Seti@Home.

Typical temparatures in hot weather were 60-65c. I've recently changed my case from the old beige plastic clad steel chassis to a lovely new Coolermaster aluminium one (see my Computer Details in my profile for details). My temps don't go over 55c anymore.

I'd say that your temps are just fine.

Ian.

[pablo180]

Cheers mate that is a weight off my mind, I was starting to get seriously worried!

For some reason I thought that it should be around 25c or something, I think that I was just getting paranoid!!

Thanks.

[bugster]

I've got both AMD and P4 systems at home and work, and by no means will the AMD systems operate as well as the P4, and that's with the EXACT SAME hardware and software configuration. The only differences maybe being in the MoBo, CPU, and maybe RAM from system to system.

Not much diference there then! [/b]

[WeedVender]

Weedvendor,
No, No, NO! It is NOT the CPU with a 1 degree per second temperature rise, it is the limit of ability of the thermistor to read more than 1 degree per second OF temperature rise. There is a world of difference there. Read the above. AMD says to operate for even a few seconds without proper cooling WILL cause failure, do not attempt to boot without heatsink and fan applied, as it WILL cause catastrophic failure.


[/u]

I stand corrected... thanks for pointing it out..

Err..
txpharoah I said that because someone kept on saying that AMD is not used in the proffesional aspect of processing. Not because of Video numbers. Donc, building CG effects and then joining it with video is the same as video editing.

[rhegedus]

Something I found about AMD failure rates:

http://www.overclockers.com/articles667/

Don't think much can be made of it though, but it might explain some people's experiences.

[txpharoah]

Something I found about AMD failure rates: http://www.overclockers.com/articles667/
Don't think much can be made of it though, but it might explain some people's experiences.

Read that. Interesting... Now consider this...

If I can buy underwear that comes with that pretty little piece of paper that says "INSPECTED BY #34" or something, why can't AMD do that with their processors? Companies are willing to spend minimum wage for somebody to check the elastic of my drawers, but another is not willing to check the structure of a product that costs dozens, if not hundreds of dollars. Says a LOT about that company, at least to me.

[WeedVender]

Your motherboard was probably made in Taiwan and probably your intel processor. Same as AMD. It doesnt matter which company... people still work as slaves.

[Colindale]

The 1% increase you *might* see with an intel cheap using SSE2 isn't worth the negative aspect of owning an Intel system and having your computer literate friends laugh at you.[/quote] ????

Boy are some of you stuck in a rut!!!!!

If you play "Games" buy AMD, if your serious buy "Intel"!

1.6 P4 running @ 2.53 rocksolid on Asus board for 15 months @33 celsius.

[LanEvo7]

Something I found about AMD failure rates:

http://www.overclockers.com/articles667/

Don't think much can be made of it though, but it might explain some people's experiences.

You people should read the whole article. If you read the whole article, note that there are MANY people that say that they do not have issues with their chips. Also if you actually read closely, it is mostly the OEM chips that are having issues. OEM chips come in bulk and are NOT packaged individually. So its not surprising that some come come with bent pins considering many suppliers do not package the cpus carefully. If you read ALL the way to then end, not just two or three lines, it says that the Retail package chips have very low failure rates. Most people that buy Intel buy it in a retail package. If you want to be safe then buy retail package.

[txpharoah]

OEM chips come in bulk and are NOT packaged individually...... If you want to be safe then buy retail package.

Why? One box is no better than the other. So, you're telling me to go pay more because some clown stamped his name on the box? Rather than buying it from an OEM's overstock? Of the same thing? Uh huh. This is another one of those urban legends.

Reminds me of the argument of why Heinz Ketchup is better than the HyTop Catsup. If you follow along carefully, you'll see both come from the same tomato paste plant. At least they did back when I was in college, and I conducted a consumer research poll on one brand being "better" than the other for superficial reasons such as this.

The entire argument of OEM vs RETAIL is a lot of hogwash that too many people, including some of the best of techs, fall into. It HAS been true, but not anywhere near as widespread as people like to make it seem. Much like how this board is full of people complaining that PRINCO IS WORST, when other boards online have the exact opposite information.

Believe what you want, but on this case, I'm afraid you are dead wrong. I have conducted much research into this topic, and written quite a few articles on it in my day. This is part of my field, and one of my primary interests: consumer myths. It is a common problem to find in the media relations business. You always have to work with people's perceptions, no matter how misinformed they may be.

[LanEvo7]

Why? One box is no better than the other. So, you're telling me to go pay more because some clown stamped his name on the box? Rather than buying it from an OEM's overstock? Of the same thing? Uh huh. This is another one of those urban legends.

Reminds me of the argument of why Heinz Ketchup is better than the HyTop Catsup. If you follow along carefully, you'll see both come from the same tomato paste plant. At least they did back when I was in college, and I conducted a consumer research poll on one brand being "better" than the other for superficial reasons such as this.

The entire argument of OEM vs RETAIL is a lot of hogwash that too many people, including some of the best of techs, fall into. It HAS been true, but not anywhere near as widespread as people like to make it seem. Much like how this board is full of people complaining that PRINCO IS WORST, when other boards online have the exact opposite information.

Believe what you want, but on this case, I'm afraid you are dead wrong. I have conducted much research into this topic, and written quite a few articles on it in my day. This is part of my field, and one of my primary interests: consumer myths. It is a common problem to find in the media relations business. You always have to work with people's perceptions, no matter how misinformed they may be.

Sure you have conducted research on this case, just like how you conducted research when you stated that AMD doesn't have SSE and the Opteron is a exclusive 64bit chip. Why do you even bother to continue to post regarding this subject when obviously you have no clue about what is going on in the computer hardware industry.

If you read the article carefully, there were MANY people that stated that their AMD's worked. The peole that reported failure gave a very vague description of why the computer failed. This includes OEM chips.

The article stated that retail chips have a much lower failure rate. The retail chip comes in its own individual package. Which includes a hsf and a three year warrenty(OEM chips typically have btwn 15-30 days). Thus eliminating the need for the reseller to package it themselves, and reduces the chance of dropping, improper packaging, etc. Also when the chips are shipped, there is more packaging on it and it reduces the chance of damage due to shocks, drops, etc. The ketchup analogy is not fit considering that we are talking about sensitive computer equipment, not tomato paste.

Regarding the Princo comment. If you are in fact correct, then why is it that in the DVD Media list, about 25% of the people report issues with Princo discs, while only about 10% have issues with Ritek. Thats just in this site, if you go over to afterdawn, there are pages talking about how bad Princo is.

I think you are the one that is spreading "consumer myths." NOT one of any of your claims has been backed by any source other than your mouth. If you think that any of my statements are false, I can immediately post up a credible source proving it.

P.S. Let clarify ANOTHER FALSE statement that you've made.

Just know that your AMD will ALWAYS be slower than a P4 on video due to the SSE and SSE2.

WRONG!

Most of the time Intel will be faster than a comparable AMD processor in video encoding, but not because of SSE or SSE2(ALL intel chips have these). There are many programs that are not even optimized for SSE(Athlon's have this) or SSE2(only Opteron's have this) and if it is, the boost is minimal. The reason the Intel's P4's perform better is because it has a higher FSB. Athlons have either a 266Mhz fsb or 333Mhz fsb, while Intel's P4s have 400Mhz fsb, 533Mhz fsb, and more currently 800mhz fsb. Video encoding is a bandwidth intensive task, so that is why P4s perform better. [/b]

[txpharoah]

On the thread:

Which goes faster with AMD? Heck, you got me. With all the new stuff that appeared in past months, who knows. But comparing the chips to Intel, from my experiences, you'll need to watch out for the heating issues when using for long periods of time and taxing themm with video work.

And RETAIL is not better than OEM. Save some bucks. Maybe use those on a better cooling system since the AMD chips can run hot without internal protection.

To LanEvo:

My research was conducted on consumer myths, not on this thread. Re-read what I wrote, and take it into context. Of them, the Ketchup one was my favorite, and back in late 2001 I conducted an OEM vs RETAIL research study. These were controlled studies, where one party was purposely told they had the opposite components. In almost all cases, consumer perception ruled where fact did not. And you keep whipping a dead horse. That's fine, I already said that. You pointed me out wrong on something. Quit patting yourself on the back. My information on SSE was old. My info on SSE2 was not too far off, with the intro of the 64-bit chips. Was I wrong on those two things? Certainly, but not because I was dumb, but merely because my information was a few months old. Most video programs take full advantage of the SSE and SSE2. That's why Intel is faster, though a better bus will surely help. Add that as another reason to get Intel. It's hard to stay on top of all the newest developments when you don't eat/breathe the stuff all the time. But the consumer myth stuff does not change next month. In fact, it will be the same in several decades as it was several decades ago. I'd love to go into the research methods, outcomes, etc. of my studies, but I normally get paid for that, and much of my stuff has already been published. Go find it on your own. I've giving you the brief: the OEM vs RETAIL is a myth in almost all cases found, with a very minute amount of truth (about 2 percent if memory serves), and with about a 2 percent margin of error. So we're looking at 96 percent guaranteed malarky that RETAIL is better than OEM. The ketchup 'analogy' absolutely fits, as did my one from the second page concerning underwear inspections. These are all products. And in all cases, it was not 'just an analogy' as much as the same situation with another product as the example. The media lists on this forum are consumer-made. Again, these are consumers that fall for a myth. Since this site is so quick to poo-poo Princo, they accept the easy answer and move on. Is there some truth to these accusations? I'm sure there is. But not some bogus number of 25 percent of all Princo failing. They'd not survive. Remember that companies buy most goods, not people. They'd have lost their corporate accounts long ago if that were the truth. Might I suggest you go to a community college and take a course in consumer psychology? Maybe grab one in mass media while you're there. Some of this stuff is 101-style information. Thanks for pointing out the info on the new AMD systems, but as I quit talking when the conversation got beyond me on the new CPUs, I suggest you do the same in this situation.

[jolo]

txpharoah ,

Your remarks are hilariously innacurate. But I wish I had your money to throw away on marketing gimicks and high prices.

There certainly are many factors about a PC besides the CPU. I'm not going to go into them, but the only thing that is a fact is that as far as price/performance and reliability go, Intel cannot touch AMD. But txpharoah, keep paying those extra dollars that help support their advertising agencies.

Even with the higher markups, higher margins offered stores, pressure to have vendors sell Intel over AMD, ignorance of people that buy only "big names" or the ones that advertise the most, AMD has crept up by 3 percent on Intel's domination in the last year.

As buyers increse their education about purchasing with money tighter, I think it will increase AMD sales. People will not just purchase Intel like automatons that they have in the past due to the marketing and vendor pressure.

It is now, what is your best bang for the buck time and in that case, not only do you get much more for your dollar with AMD, the mother boards are less money as well.

[tww1]

txpharoah you can say all you want to about the AMD Chip been runing them now for two years made around 1000 movies and not a burn down problem never a shut down and time from the P4 the AMD beat it hands down for compressing files of the same size useing the same MB and same HD's , 512Ram in both P4 and the AMD. Converting a 1Hr avi file in TEMPG to SVCD took the P4 2.4 1:45 min the AMD XP1700 1:15 min. The only question i have is why would i want a P4 to do video?

[LanEvo7]

What are you trying to do? Making the font extra small so I can't point out your mistakes?

My information on SSE was old. My info on SSE2 was not too far off, with the intro of the 64-bit chips. Was I wrong on those two things? Certainly, but not because I was dumb, but merely because my information was a few months old.

Like I saide before, not a few months, you are about a two years behind. With computers components changing so rapidly, your information is VERY aged. People come here for advice so its not nice to give people outdated information. Not only have I been pointing out your untruth statements(outdated info, whatever you want to call it) other posters have too. ANd you still haven't shown me a bit of proof to back of your other "facts".

Go do some research, bandwidth contributes a lot more than SSE or SSE2, but of course you won't bother and continue this ignorance.

I'm done with psy and those are theories only. There are a lot of them and some even contradict each other. But I won't go into that.
Lets just take your theory for example, according to your theory, AMD's only have a bad rep because of people like you. You bad mouth them, then when people buy a system with the AMD processor in it or build one themselves, at the first sign of failure they blame the manufacturer.

If according to you that their processors fail so often, they why does their company continue to grow? How come IBM created a partnership with them to create a new processor? They should be bankrupted by now. [/quote]

[gmatov]

Actually, I think you can make a lot out of it. Basically, the site is telling you that AMD failure is anectdotal, someone knows someone whose CPUs failed, and, more'n likely, too many people know that poor slob and all reported it to sites such as this. All of a sudden, they're dropping like flies.
I have mentioned that I have only had 1 actually burn out on me, and I'm not sure it wasn't the cheap all-in-one MOBO's fault. It was a 4th unit, a 1 gig, soon to have been handed off to one of my g'kids. Burned up, first, he wound up with a 1.2, and a better MOBO, another 1.2, going to the g'daughter, soon as I can get a DVD -player at the next show, 2 weeks away. Hey, she's 11. 1.2 should be good for middle school.
I wonder how prevalent counterfeits are in AMD world, laser re-etched, if the new buyer had time and inclination to run AMD's CPUID, and if it didn't go out soon after boot-up, maybe some of them are burning up 1 gigs re-marked as 1.6s. That supposedly was a big problem in years past, perhaps still is?
I'd like to see someone point out a site with anectdotal evidence of Pentium burnouts. You know they're out there. Nothing is perfect, and I did mention that we buy CPUs that have failed all the tests above the level at which they are marked and marketed as. It's also entirely possible that Intel marks theirs 2 bins down to keep from having burnouts.
If you stay a step or 3 down from the leading edge, they're too cheap to complain too loudly about. Unless it's Intel, of course.