"Real World" CPU Test at THG

[AnandThenMan]

I have no problem using the stock hardware that comes boxed with the processor. But why run everything in the open air? This is USELESS. It tells you nothing about how the system will actually perform in a closed space. Let's face it, the entire test is flawed.

 

[jiffylube1024]

OK, so another fun review from Tom's Hardware where they don't interpret the data in any meaningful way. First of all they don't comment on throttling at all for the Intel (I'll assume it's not throttling because if you check the temperatures, they aren't hitting the 80C 'throttle limit' even with +15C). More importantly the Vcore isn't dropping by margins of .1V or greater at any points so it's safe to say the chip isn't throttling.

Temperatures between the two are not comparable to one another: they are using vastly different stock cooling of the two chips; the AMD stock cooler is a POS (much like how ATI has cheap coolers on the X800 cards. Why? Because their stuff runs cooler so they can afford to cheap out on the cooling and save money that way).

We already know that the Prescott consumes considerably more power than even the 3800+ chips (and up), so this temperature comparison is only valid for what THG intended: a comparison of the two systems using stock cooling.

If you want an accurate comparison of operating temperatures, first you'd be using the same HSF for both and second of all you'd be using an external, calibrated temperature monitor, not relying on the CPU/Motherboard to interpret it for you.

So, no Winchester is not "hotter" than Prescott, but Winchester's stock crappy cooling is even worse than Intel's crappy stock cooling (personally I would buy a better HSF for either one, especially with the heat they output).

Tom's 'comparison' is nothing more than two separate tests to see if either architecture becomes unstable at 100% load using stock cooling. However he further confuses the matter by doing the tests in an open-air controlled environment.

Furthermore, he's using the second fastest Intel chip (3.6 GHz) while using a chip 3-4 speed grades from the top on the AMD side (3500+). You'd think the logical comparison would be the 3.6 Ghz (second fastest) with the 3800+. There would most probably be differences in the stock cooling between the 3500+ and 3800+ but oh well... This is how THG likes to do it: a random pairing of chips.

Don't try to infer generalizations about temperature or comparing architectures from these tests. The only meaningful results we can get is whether or not either system becomes unstable, crashes or throttles.

 

[AWhackWhiteBoy]

This test is a question of stability and heat isn't even introduced to the setup as a whole, I suppose Tom didn't want to see the Intel throttle at all...

 

[thegimp03]

lol...

 

[SonicIce]

What about noise?

 

[biostud]

Originally posted by: klah

Originally posted by: biostud
The same cooler on both processors would have been more fair I think.

If the purpose of the article was to compare temperatures perhaps. But the purpose is to test retail parts, assembled with the stock hardware. If the retail Prescott comes with a larger, more powerful HSF then it is appropriate to use it in this instance.

Then the temperature gauge doesn't really matter that much. It just tells that Intel needs a better fan to cool their chips to the same temperature.

But if the test "just" is to test up-time of retail parts, then it seems ok. It isn't very helpfull for the PC enthusiast though as we would have chosen different parts.

 

[Rand]

Originally posted by: Mik3y
tom is pretty biased towards intel, and this isnt the first time where he's bent the better odds towards intel products.

I'm not all that fond of Tom Pabst attitude, and I often find is choice in benchmarks.... peculiar to say the least but biased towards Intel?

My how quickly people forget. I can recall some awfully harsh rants he threw out in the era of the much maligned PIII Coppermine 1.13Ghz, and to a lesser degree the early Willamette in respect to it's relatively lackluster IPC.
Let us not comment on is thoughts with respect to the Willamette based Celeron, DRDRAM mated with the P6 core, or SDR SDRAM with the P4.

 

[classy]

Tom has been controversial over the years, but Tom's is still one of the top 2-3 hardware websites on the net. Even if you don't care too much for the head person, Tom, his website turns out some the best stuff on the net. The test is that of stress, not how much heat each is producing. Granted I noticed that as well that the AMD by their numbers is running pretty toasty. But as of this morning, the AMD is still chugging along and the Intel machine has rebooted. I think this is a pretty neat test.

 

[FlameDeer]

Under stress, P4 PressHot already restart once few hours ago. In the meantime, Athlon 64 still going well.

 

[Acanthus]

Im running a prescott 3.0 @ 3.73ghz and stress testing it right now, and its not shooting flames, its not exessively loud, and its not throttling...

On a stock cooler, at stock voltage.

All the intel hate needs to stop already

 

[Duvie]

Wingz...

I have an intel northwood 2.4@3.5ghz with a kick-arse cooler and it loads to 58-60c on my abit mobo....My AMD 3000+ with a similar all aluminum stock cooler with low cfm fan on Toms MSI board did 56-58c load at 2.6ghz...each with vcore boost up near 3-4%....

The facts are hard to see here... the temps of mobos are widely suspect, but that being said I couldn't run the 2.4 with a stock cooler....Lets just temps hit near 66-68c with an alpa still better then stock cooling...


The fact I know is I am running with a venus TT now with 3 of my 5 case fans turned off and only loads at 50c....real world is the 90nm winnie is much cooler...

Edit: and we all know the prescotts (common stepping) runs hotter clock for clock against the northwood, so their is no way I would have the same cooling eefct at 3.5ghz with a boost in vcore...probably be well within the mid 60's and that is with an aftermarket cooler....

Tom is suspect, but read between the lines in his articles and other forums and you can see through the BS....I always have even when I ran my northwood...

 

[Nyati13]

Originally posted by: FlameDeer
Under stress, P4 PressHot already restart once few hours ago. In the meantime, Athlon 64 still going well.

If you look closely, you can see that the temps of the AMD system come down a bit, right during the time that the Intel system is down. So the AMD temps are being effected by the Prescot running next to it.

 

[Zebo]

Originally posted by: Nyati13

Originally posted by: FlameDeer
Under stress, P4 PressHot already restart once few hours ago. In the meantime, Athlon 64 still going well.

If you look closely, you can see that the temps of the AMD system come down a bit, right during the time that the Intel system is down. So the AMD temps are being effected by the Prescot running next to it.

Originally posted by: Acanthus
Im running a prescott 3.0 @ 3.73ghz and stress testing it right now, and its not shooting flames, its not exessively loud, and its not throttling...

On a stock cooler, at stock voltage.

All the intel hate needs to stop already

It's winter time. Why you think you're seeing these great clocks at 2am from users on both amd and Intel?. It's cold!!!.. Wait till it hits 90F under a desk with dust bunnies in the summer at joes sixpacks house.

 

[PingSpike]

This seems like a horribly unscientific test at best, since they've done little to eliminate outside variables (motherboard temps are inaccurate, stock cooling solutions that are different can be argued against) and have even gone so far as to introduce some. (Why are they running right next to eachother? There's no good reason.)

Given that Tom's has gained a reputation for being a bunch of Intel fanboys as of late, I would think they would have been more careful about a test this controversial. The test reeks of Intel bias IMO, and is really forcing me to view toms as trash. Because thats what bias reviews are, trash. No matter who they are bias against.

But perhaps I'm being a bit harsh. It could just be that they're stupid.

 

[Zap]

Originally posted by: klahFor example the latest bios for that MSI board reduced the temperatures reported by 10C, presumably so that customers would stop whining.

I have an Abit IS7... didn't Abit do that as well with a BIOS update?

 

[gururu]

OMG! Prescott doesn't run much hotter than an an AMD cpu! Anyone with sense would have realized by now that Prescott's aren't nuclear reactors and DON'T melt down.
OMG! AMD actually puts up some warm numbers! Go figure.

Can we please put a rest to the temp issue, because it is a non-issue.

 

[Duvie]

Originally posted by: gururu
OMG! Prescott doesn't run much hotter than an an AMD cpu! Anyone with sense would have realized by now that Prescott's aren't nuclear reactors and DON'T melt down.
OMG! AMD actually puts up some warm numbers! Go figure.

Can we please put a rest to the temp issue, because it is a non-issue.

Not for persons with C Northwood chips such as ourselves....BUt a 2.8 prescott placed in same mobo with same hsf you have now and clocked to 3.5ghz with similar vcore boost is warmer and that is a fact borught to me by many Intel users. This is also why many Intel fans have either tried to get higher sp[eed P4cs or have skipped the current prescotts all together.


I say that mobo temp reporting is so bad that it is hard to get reliable numbers, period.....

I just can tell you with a certainty my room is cooler, the 2 case fans I still have running out of my 5 originals with my p4c are pumping out less heat from the case, and my temps reported (take with a grain of salt) are less with a smaller 80mm (less cfm) fan.

the overall interior of the case environment is cooler. My power supply seems cooler and I have even turned the fan down to minimum on that as well....

It is undeniable to me that this(A64 90nm 3000+@2.6ghz) is cooler then my p4c northwood and thus all reports of northwood to prescott reviews have shown that the northwoods are cooler then the prescotts by anywhere from a few C to a decent amount....


Edit: By the way all of my components are the same except the cpu/mobo in my A64 setup....same vid card, same power supply, same case, same drives, etc....

 

[FlameDeer]

Ops, Intel P4 reboot again around 30mins ago! :shocked:
Notice that Athlon 64 system cpu temp & northbridge fan speed drop every time when P4 is offline.
This show clearly Athlon 64 place beside P4 really affected by high temp & high power consumption of P4 prescott.

 

[superkdogg]

They did an edit stating that the two systems were going into high-quality cases to create a more real-world scenario. I don't see any cases, what's up with that? And I was skeptical of the AMD being affected by the P4's heat-right up until looking at the NB fan tracking. It dropped DRAMATICALLY when the P4 was down. That is not coincidence and is presumably not error, that leaves only a causal relationship. Anyhow, Intel fanboys (I own both and root for neither) it is a fact as well as a generally held truth that Prescotts are hot and A64's are not as hot. Check out the max wattage of a winnie and that of Prescott-it's nearly 1/2! How could that not mean 'cooler?' Regardless of all the hoopla, not setting up independent temp. monitoring and also not tracking ambient temp. is amateur hour. I bet they're thinking about ways to screw up future tests as we speak!

 

[MDme]

I was right!!! damn open air test!!!! the prescott is heating up the AMD cpu. the fact that the intel system is rebooting might probably be second to the heat it is sustaining. the AMD CPU might actually be cooler than what the temp sensor shows that's why it's stable.

 

[Shaotai]

Originally posted by: Nyati13

Originally posted by: FlameDeer
Under stress, P4 PressHot already restart once few hours ago. In the meantime, Athlon 64 still going well.

If you look closely, you can see that the temps of the AMD system come down a bit, right during the time that the Intel system is down. So the AMD temps are being effected by the Prescot running next to it.

BUWAHAHAHAHAHA!

(My vote for the funniest quote of this thread...)

And Why did the temp of the intel syste jump 10 degrees! Am I looking at it wrong?? Yikes!!

EDIT - More like 15 degrees... wow... maybe we'll see a screen shot of it going up in flames shortly...

 

[superkdogg]

Temps are up on the P4 because the 12v rail is also up. Are they upping the voltages to try to keep the P4 from crashing, or am I a conspiracy theorist?

 

[FlameDeer]

After 2nd reboots, by increase of +12v & +3.3v, P4 system is now reaching a very high CPU temp of max 76C in Munich winter night.

 

[MDme]

hey, where's the 76C temp you mentioned? wanted to see it (and maybe see the P4 in flames too) lol

 

[jiffylube1024]

Originally posted by: MDme
hey, where's the 76C temp you mentioned? wanted to see it (and maybe see the P4 in flames too) lol

Follow your own link to the Intel setup!

The Intel is much hotter after the reboot for some reason - it's up around 75C all the time! Prescott doesn't have a heat problem, eh...?

Also, the Intel system is going haywire - despite temperatures the fan speed is lower than before the reboot. Voltages are about +0.1 higher after the reboot but Vcore is actually lower and jumping around more! That P4 is going to kill the PSU or burn out the mobo, I swear!

(Is this all a cherade by Tom's to create more attention and drama? What changes took place during the reboot to cause the funky voltages, jumpy Vcore and lower fan speed?)