Compare your CPU to a 4Ghz Core i7 965 EE Nehalem!

[Tempered81]

Here's a screenie of a 4000mhz Core i7 965 doing a 16m SuperPi run. It finished the benchmark in 4min 46sec. My 1800mhz merom took about 12 minutes - about 2.5 times as long.
Anyone out there with a 4Ghz cpu want to try it out and compare to a 4Ghz i7 965 EE? We should be able to see a nice clock-for-clock comparison with the results.

Download superPI here (make sure to do the 16m test):
http://www.xtremesystems.com/pi/super_pi_mod-1.5.zip

Compare to this screenshot of Nehalem Bench (4 Minutes 46 Seconds):
http://img216.imageshack.us/my.php?image=qpi210nf9.png

Make sure not to run any other tasks/processes while performing the benchmark.


Core i7 965 EE @ 4ghz: 4 min 46 sec

-EDIT-
SuperPi 16M results from people in this thread:

Core i7 920 @ 4.5Ghz: ``````` 3 min 38 sec
Core i7 920 @ 4.0Ghz: ``````` 4 min 10 sec
Core i7 940 @ 4.2ghz (4.0 Trb): 4 min 13 sec
Core i7 920 @ 4.0Ghz: ``````` 4 min 16 sec
Core i7 965 XE @ xxGhz: ````` 4 min 18 sec
Core i7 920 @ 4.0Ghz: ``````` 4 min 24 sec
Core i7 965 XE @ 4Ghz: `````` 4 min 46 sec
Core i7 920 @ 3.7ghz: ``````` 4 min 52 sec
Core i7 920 @ 3.4ghz (3.6 Trb): 4 min 52 sec
Core i7 920 @ 3.4ghz (3.6 QPI): 4 min 55 sec

QX9650 @ 4.0ghz: `````````` 4 min 52 sec
Q9550 @ 4.03ghz: `````````` 5 min 35 sec
Q9650 @ 4.05ghz: `````````` 5 min 53 sec
Q9650 @ 3.6ghz: ````````````6 min 55 sec
Q9450 @ 3.6ghz: ````````````6 min 12 sec
Q9300 @ 3.45ghz: ```````````6 min 25 sec
Q6600 @ 3.6Ghz: ````````````6 min 27 sec
X3350 @ 3.4ghz: ````````````6 min 29 sec
Q6600 @ 3.68ghz: ```````````6 min 55 sec
Q9450 @ 3.4Ghz: ````````````6 min 34 sec
Q6600 @ 3.3ghz: ````````````7 min 29 sec
Q8200 @ 3.0Ghz: ````````````7 min 34 sec
QX9650 @ 3.0ghz: ```````````7 min 39 sec
Q6600 @ 3.2ghz: ````````````7 min 44 sec
Q6600 @ 3.3ghz: ````````````8 min 10 sec
Q6600 @ 3.2Ghz: ````````````8 min 55 sec
Q8200 @ 2.0Ghz: ```````````11 min 04 sec

E8600 @ 6.2ghz: ````````````3 min 35 sec
E8xxx @ 4.6ghz: ````````````4 min 46 sec
E8xxx @ 4.2ghz: ````````````5 min 33 sec
E8xxx @ 4.2ghz: ````````````5 min 38 sec
E8xxx @ 4ghz: `````````````5 min 40 sec
E8xxx @ 4.1ghz:```````````` 5 min 43 sec
E8xxx @ 4.2ghz: ````````````5 min 44 sec
E3110 @ 4ghz: ``````````````5 min 48 sec
E8xxx @ 4ghz: ``````````````5 min 59 sec
E8xxx @ 4ghz: ``````````````5 min 58 sec
E8xxx @ 3.8ghz: ````````````5 min 59 sec
E8xxx @ 4ghz: ``````````````6 min 06 sec
E8xxx @ 4ghz: ``````````````6 min 08 sec
E8xxx @ 4ghz: ``````````````6 min 25 sec
E8xxx @ 3.6ghz: ````````````6 min 36 sec

E7200 @ 3.6Ghz: ````````````6 min 55 sec
E6750 @ 3.4Ghz: ````````````7 min 14 sec
E5200 @ 3.25Ghz: ```````````7 min 26 sec
P8xxx @ 2.53ghz: ```````````8 min 46 sec
E7200 @ 2.5Ghz: ````````````8 min 48 sec
E2180 @ 3.0ghz: ````````````9 min 02 sec
E2180 @ 3.4ghz: ````````````9 min 28 sec
Athlon 64 X2 @ 3.08ghz: `````11 min 55 sec
T7100 @ 1.8ghz: ```````````13 min 11 sec
T2500 @ 2.0ghz:````````````13 min 38 sec
Pent 4 650 @ 3.4ghz:````````16 min 19 sec
Pent IV @ 2.91Ghz: `````````17 min 55 sec
Turion @ 2.1ghz: ```````````18 min 35 sec
Athlon 64 @ 2.0ghz: ````````18 min 49 sec
Pent IV @ 3.0gHz: ``````````20 min 46 sec
Pent M 735 @ 1.8ghz: ```````24 min 03 sec
Pent M @ 1.4ghz: ```````````26 min 57 sec
Pent III @ 800mhz: `````````63 min 19 sec

^^^Single threaded PI performance delta i7 vs. E8xxx based on above results: ~26%

Single threaded PI performance delta i7 vs. QX9770 based on review here: ~19%
http://xtreview.com/addcomment...re-i-7-940-review.html

 

[Markfw]

8 minutes 5 seconds for a 4050 mhz E8400

 

[Tempered81]

Thanks! seems it's 40% faster clock-for-clock in this case, mark. Anyone out there have a 4ghz quad? e5200, e7200, phenoms?

 

[AmberClad]

Originally posted by: jaredpace
My 1800mhz merom took about 12 minutes

Originally posted by: Markfw900
8 minutes 5 seconds for a 4050 mhz E8400

*scratches head*

Maybe I need some sleep, but those two sets of results don't jive...

 

[Markfw]

Originally posted by: AmberClad

Originally posted by: jaredpace
My 1800mhz merom took about 12 minutes

Originally posted by: Markfw900
8 minutes 5 seconds for a 4050 mhz E8400

*scratches head*

Maybe I need some sleep, but those two sets of results don't jive...

Yup, I agree. But my 8 minutes vs ~5 for the i7 seems about right.

 

[rge]

My E8400 OCed to 4 ghz and running ram 1:1 9x445 runs superpi 16m in 5 minutes 40 seconds.
http://img366.imageshack.us/my...age=superpi16m2dg5.jpg

 

[Cogman]

18m 49.994s on a Athlon 64 3200+ at 2.0 GHz. Assuming a linear scale that puts it at ~ 9m 25s at 4.0 GHz

 

[disports]

6 minutes 25 seconds for my E8400 at 4005 MHz

 

[Dopekitten]

With a Q9450 @ 3.6ghz i get:
6min 12secs

Im only 1min 26s behind, and im not even at 4ghz. So im not really sure if those "30% faster" clock for clock claims have any base...

The only thing i see in Nehalem is greater cache and hyperthreading, but its not like i need anymore than 4 cores right now .

Edit: Calculated it out. If it scales linearly, i should get around 5m 20s at 4ghz, and should need around 4.4ghz to match the Core i7 965 @ 4ghz. So i guess it is faster, but this puts it at only a mere 10% faster at single core applications.

 

[Comdrpopnfresh]

Not to butt heads, but I disagree with this comparison point. I think there is a reason why 1m tests in superpi that have been used as an architecture to architecture comparison. Anything less doesn't allow for improvements to show, and anything more allows optimizations to contextually bias the test: kind of like power utilization tests are often done with c1, powerstep, q'n'c disabled.
That being said, I don't have a nehalem system, and if that generation of microprocessors is so fast that it chews through floating point numbers as quickly as a c2d can go through 64k computations of pi, then I'm all for a higher computational load than 1m being used. But if it is, then it cannot be used as a basis of comparison between nehalem and c2d. You can't compare a pII with a c2d based on 1m pi tests- it's somewhat absurd, because the difference is just so large. If a cpu is a revolutionary advance, that leads previous evolutionary designs in the dust, than that nature should just be accepted as well as the knowledge that comparison isn't necessary.
Maybe 2m or 4m tests?
Then again, you are comparing 'systems' vs processing...

 

[AmberClad]

Ugh...

All I can say is that Super Pi does not like Vista. If I run Super Pi at all, I usually do it on my XP hard drive which has all of my other stress test tools. Unfortunately, that would involve finding and installing that HDD, so here are the Vista results. I had to disable Windows Audio and Aeroglass to get it to even run :roll:. On top of that, Super Pi also runs a bit slower on Vista, so it's not an apples-to-apples comparison since that screenshot in the OP seems to show XP.

But I think we might conclude that Mark had something else running while he ran the test (like Folding, maybe ?).

 

[dbcooper1]

6 minutes 8 seconds on my E8500 at 4GHz

 

[Dopekitten]

@Amber, i am running vista, see my above post for what i got. It's still a valid comparison i believe.

 

[Tempered81]

Originally posted by: Markfw900

Originally posted by: AmberClad

Originally posted by: jaredpace
My 1800mhz merom took about 12 minutes

Originally posted by: Markfw900
8 minutes 5 seconds for a 4050 mhz E8400

*scratches head*

Maybe I need some sleep, but those two sets of results don't jive...

Yup, I agree. But my 8 minutes vs ~5 for the i7 seems about right.

well your 8 minutes is a bit slower than the other four e8400's here that are at around 6 minutes.

Originally posted by: AmberClad
But I think we might conclude that Mark had something else running while he ran the test (like Folding, maybe ?).

 

[Cheex]

Just to let you know...when my E7200 comes this week, I should have it at 4GHz by the weekend, I'll chime in with my results.

 

[Idontcare]

My Q6600 @ 3.3GHz does 16m SPi in 7m 39s.

My P3 @ 0.8GHz does 16m SPi in 63m 19s.

Originally posted by: Markfw900
8 minutes 5 seconds for a 4050 mhz E8400

Mark did your BIOS lose your OC on ya during a reboot? Compared to my G0 score, an 8min E8400 score sounds like a stock E8400 super-pi run.

 

[AmberClad]

Even among the 4.0GHz results, there's a decent amount of spread, but that's to be expected. Different people obviously have different things running in the background. In my case, I'm got 63 background processes. Someone using a system primarily for benching is probably going to be running a much more "bare" configuration.

So, overall it's not exactly a scientific survey.

I think the worst-case-scenario that Anand saw was a 2% improvement over a Penryn, and that was single-threaded Cinebench (but that was with a system that was still in "beta" hardware). Super Pi is of course single-threaded, so it's not going to play to i7's strong suits.

 

[Markfw]

Originally posted by: Idontcare
My Q6600 @ 3.3GHz does 16m SPi in 7m 39s.

My P3 @ 0.8GHz does 16m SPi in 63m 19s.

Originally posted by: Markfw900
8 minutes 5 seconds for a 4050 mhz E8400

Mark did your BIOS lose your OC on ya during a reboot? Compared to my G0 score, an 8min E8400 score sounds like a stock E8400 super-pi run.

Well, you are right, its looks that way, but cpu-z says 4050, and I was running F@H, but I shut it down and reran with nothing, same time. But its win2k, maybe thats the problem.

 

[Sureshot324]

Interesting that all kinds of benchmarks have been 'leaked' but we haven't seen a single game benchmark.

 

[Cheex]

Originally posted by: Sureshot324
Interesting that all kinds of benchmarks have been 'leaked' but we haven't seen a single game benchmark.

Well it has been said that Core i7 won't benefit gamers much...so...maybe that's why they are "hiding" those benchmarks.

 

[Markfw]

I just tried it on a 3366 Q6600 using XP, and got 8 minutes 10 seconds. I think XP works better than win2k.

 

[disports]

I just ran it again with nothing else running and it was 5 minutes 58 seconds. 27 seconds faster lol.

 

[Enthusiast]

E8600 @ 4.6GHz - 4m 46.931s

http://i431.photobucket.com/al...33/CrossFireX6/16m.jpg

 

[apoppin]

Originally posted by: Cheex

Originally posted by: Sureshot324
Interesting that all kinds of benchmarks have been 'leaked' but we haven't seen a single game benchmark.

Well it has been said that Core i7 won't benefit gamers much...so...maybe that's why they are "hiding" those benchmarks.

haven't we seen them?

they suck

Am i supposed give a crap about a a 16mb SuperPi run?

 

[Tempered81]

It's just a pre-cursor to the gaming benchmarks, apoppin. That's all.