My Intel Core i7 and Phenom II 940 Comparison

[Lionheart]

just muckin around, by the way man, you got 2 awesome rigs!

You wanna swap, lol!

 

[Chicken Patty]

just muckin around, by the way man, you got 2 awesome rigs!

You wanna swap, lol!

no thanks.

 

[Tatty_Two]

did you guys know that AMD rulez, lol

Sniff Sniff...................

 

[lemonadesoda]

Two things for performance comparison.

1./ HT should be ON, IMO, otherwise you are NOT showing what the chip is capable of
2./ You need to set threadcount in wPrime to 32. The default of 4 FAILS to show true performance of a 6, 8, 12, 16 "core" or "thread" CPU

=======================================

I've done some comparisons for you on Dual Xeon E5420 STOCK (ie 2.5Ghz, 8 real cores)

wPrime = 9.203 secs = (= 23.000 secs divided by Ghz)
CPU score on 3Dmark06 = 5636 points (= 2250 per Ghz)

 

[Binge]

Two things for performance comparison.

1./ HT should be ON, IMO, otherwise you are NOT showing what the chip is capable of
2./ You need to set threadcount in wPrime to 32. The default of 4 FAILS to show true performance of a 6, 8, 12, 16 "core" or "thread" CPU

=======================================

I've done some comparisons for you on Dual Xeon E5420 STOCK (ie 2.5Ghz, 8 real cores)

wPrime = 9.203 secs = (= 23.000 secs divided by Ghz)
CPU score on 3Dmark06 = 5636 points (= 2250 per Ghz)

That's very short sighted. I have posted before in defense of his testing methodology but here it comes again. HT is OFF because he is comparing the cores. We all know what HT does, but when it just comes down to which chip is the better quad core clock for clock then it shows a different kind of strength than HT VS non-HT. That to me and every sane person is more of an apples vs oranges instead of coke vs pepsi comparison.

HT off is also how a number of users will get a 4.0GHz overclock with the i7 because not every chip is capable of 4.0+ with HT on. Turning HT off can also give more life to the processor by reducing temperatures. To be blunt... it's short sighted of you to expect CP to give people information they already have. It's much better he gives them a practical view of both chips instead of info people can find in run of the mill reviews.

 

[Chicken Patty]

That's very short sighted. I have posted before in defense of his testing methodology but here it comes again. HT is OFF because he is comparing the cores. We all know what HT does, but when it just comes down to which chip is the better quad core clock for clock then it shows a different kind of strength than HT VS non-HT. That to me and every sane person is more of an apples vs oranges instead of coke vs pepsi comparison.

HT off is also how a number of users will get a 4.0GHz overclock with the i7 because not every chip is capable of 4.0+ with HT on. Turning HT off can also give more life to the processor by reducing temperatures. To be blunt... it's short sighted of you to expect CP to give people information they already have. It's much better he gives them a practical view of both chips instead of info people can find in run of the mill reviews.

THank you Binge it couldn't have been said better.

However, we all know that with HT on the comparison is lop sodded. Thats why I did it with HT off to see what CPU is really quicker clock for clock. If I turn HT on, then everybody will complain as the results are not pretty for the AMD if that was the case.

 

[lemonadesoda]

Well, you could have said it in a less rude way.

If you want to compare THREAD performance (per thread) then yes, turn off HT. But if you want to build a one chip PC and you are deciding based on performance ability, then to suggest handicapping one processor is the right way of doing it, is, IMO, asinine.

Well now that I own both CPU's I figured I'd do a comparison between the both. I am only doing 3dmark CPU tests, and wprime comparisons. I dont have much time on my hands to do more benchmarks and these are the ones I have my hands on now. So lets see.

CP did not say he was running a "one hand tied behind the back" test. All I did was chip in some comparable performance stats. Only trying to be helpful. Gee.

PS. Why is CP running three channel DDR3 against two channel DDR2? If you want apples vs. apples, then for heavens sake, pull a stick out of the i7 system.

THank you Binge it couldn't have been said better.

Oh, great. What is this? The Bert and Mary show? Super-frigging-califragilisticexpialidocious.

PPS. Run the wPrime benchmarks again with a higher thread count. Check your answers.

 

[Chicken Patty]

Well, you could have said it in a less rude way.

If you want to compare THREAD performance (per thread) then yes, turn off HT. But if you want to build a one chip PC and you are deciding based on performance ability, then to suggest handicapping one processor is the right way of doing it, is, IMO, asinine.



CP did not say he was running a "one hand tied behind the back" test. All I did was chip in some comparable performance stats. Only trying to be helpful. Gee.

PS. Why is CP running three channel DDR3 against two channel DDR2? If you want apples vs. apples, then for heavens sake, pull a stick out of the i7 system.

Oh, great. What is this? The Bert and Mary show? Super-frigging-califragilisticexpialidocious.

PPS. Run the wPrime benchmarks again with a higher thread count. Check your answers.

I will say only one thing. Your post was ridiculous. Sorry but true. As far as DDR3, the Phenom II rig is running on XP while my i7 is running on W7. THe advantages of the DDR3 are taken away by using a much faster OS on the AMD rig. IF you havent seen the difference I suggest you try it out yourself.

 

[douglatins]

I would like to see real world benchmarking, IMO Crysis... Hehe games asktualy. With same GPU config

 

[Chicken Patty]

I would like to see real world benchmarking, IMO Crysis... Hehe games asktualy. With same GPU config

I will try to do some 3dmark runs with my video card in the AMD computer. However haven't had much time and I gotta take apart the rig and swap cards and stuff.

I'll try to get some more real life benchmarks in but I don't promise anything.

When I say do some 3dmark runs, is just to compare sm1 and sm2 scores, see how each CPU works with the card.

 

[lemonadesoda]

CP, before you "unbuild" your rigs, could you kindly run the following. They are EXTREMELY quick to do. It will take you less than 30 secs per bench.

1./ GeekBench
2./ Fritzchess bench
3./ Everest 5 CPu PhotoWorxx bench

Thank you kindly!

PS. If you want to see if the memory setup if "fair" across machines, use

1./ Sciencemark Membench, or
2./ Everest Memory and Cache benchmark tool

Just check the "main memory" bandwidths to see if there is comparable performance. If the results are way off due to tripple channel, then the Intel gets an "unfair" advantage due to cache pre-fetchers working much faster... which has nothing to do with the core technology, on the tripple channel... which I think you wanted to exclude from your analysis.

 

[Chicken Patty]

CP, before you "unbuild" your rigs, could you kindly run the following. They are EXTREMELY quick to do. It will take you less than 30 secs per bench.

1./ GeekBench
2./ Fritzchess bench
3./ Everest 5 CPu PhotoWorxx bench

Thank you kindly!

PS. If you want to see if the memory setup if "fair" across machines, use

1./ Sciencemark Membench, or
2./ Everest Memory and Cache benchmark tool

Just check the "main memory" bandwidths to see if there is comparable performance. If the results are way off due to tripple channel, then the Intel gets an "unfair" advantage due to cache pre-fetchers working much faster... which has nothing to do with the core technology, on the tripple channel... which I think you wanted to exclude from your analysis.

ill run them tonight. Can you please be so kind and just link me to benchmarks 1 and 2 please. I have everest already.

 

[n-ster]

I feel some tension around here xD

anyways, great benchmarks CP!

here:
GeekBench
Fritzchess

Not sure if Fritz is the right version but oh well, just trying to help

 

[lemonadesoda]

ill run them tonight. Can you please be so kind and just link me to benchmarks 1 and 2 please. I have everest already.

Careful with EVEREST - They have changed the benchmarking "code" due to i7 compatibility. Results from v4 are not valid for v5. Please use v5 Everest.

 

[btarunr]

Fair comparison would be to test the platforms in the best they have to offer, in identical software environments.

HTT should be enabled for the Core i7, it should be given 3 or 6 GB of DDR3-1333 MHz memory.

The Phenom II X4 940 should be given 2 or 4 GB of DDR2-1066 MHz memory, as that's the fastest processor by AMD at the moment, and that's the highest memory standard it is specified to run.

Both platforms should run on the same OS (with the same Service pack (if any), updates, etc.)

In pursuit of making things fair, you're giving Phenom II the unfair advantage by crippling the Core i7's features (disabling HTT) and making it handle a resource-heavier OS.

 

[lemonadesoda]

^^ all quite correct btarunr, but these guys are trying to do a "core-core" as STEP ONE of their scientific approach. Then, STEP TWO, just add the thrid channel of DDR3 and turn on the HT to watch the nitrox effect of i7 architecture.

 

[btarunr]

I jumped from post 1 to 54 and my post. I had expected some more clarity in post 1, which typically indexes the discussion. Doing a step by step comparison is fine, though it still does not warrant giving the two dissimilar software environments.

 

[trt740]

This just shows how well AMD did with these chips using existing DDR2 and existing motherboard tech. If you have a older system from AMD this is a hell of a upgrade. However, intel also came through with new cutting edge stuff so it's a win win.

 

[Chicken Patty]

here is some quick runs with the benchmarks that lemonadesoda requested. This is my daily settings of my i7 rig. Just for now so you guys have something to look at.

I'll do the comparison with HT off agaist the Phenom II tonight.

 

[n-ster]

Thanks a lot for those I was wondering what those scores would be with YOUR daily setting!

 

[Damian^]

in case you didnt notice HT was off on the intel, just in case. with HT on it wouldn't be fair.

Why not? When it comes to testing/comparing its best to test as is especially if its clock for clock. Your results are a bit misleading.

The average user normally has HT on. But i guess on special occasions if one does turn it off then thank you for the time you took to do this comparison.

EDIT: Ehh looks like some people have said what i thought too, I usually just discuss with OP not others my bad.

 

[Chicken Patty]

Why not? When it comes to testing/comparing its best to test as is especially if its clock for clock. Your results are a bit misleading.

The average user normally has HT on. But i guess on special occasions if one does turn it off then thank you for the time you took to do this comparison.

EDIT: Ehh looks like some people have said what i thought too, I usually just discuss with OP not others my bad.

you usually discuss with OP? Whats that?


Well my comparison is not what each CPU can do to its max, just clock for clock with the Phenom. If I have HT on, everybody will complain that the intel beat it because HT was on. So I disabled it.

This is not a test about what the i7 can do, for that we have the i7 discussion thread where you post your results and such. If the AMD doesn't have Hyper Threading, I don't think it is fair that HT should be on when benching the i7 against the Phenom II.

 

[lemonadesoda]

Wow. That i7 machine SMOKES those benchmarks. Very nice. I'm very interested to see how HT affects those results.

I would be surprised if Il Phenom comes anywhere close... but let's see!

 

[Chicken Patty]

Wow. That i7 machine SMOKES those benchmarks. Very nice. I'm very interested to see how HT affects those results.

I would be surprised if Il Phenom comes anywhere close... but let's see!

Sorry bro, I couldnt do it last night.

I've been in and out of hospital with health issues lately bro, lets say the past two or three weeks. I'll get those results for you soon bro. While i'm at it i'll also do the encoding benchmark 3dsage I believe wanted me to do. STill owe him that.

 

[lemonadesoda]

Take it easy. Health should always come first.