Nuclear Multi-Core CPU benchmark

[Cheex]

E6300 @ 3.15GHz (450x7)

Score: 8576

 

[Cheex]

Intel Core 2 Duo E6300 Conroe (L2) @ 3.360GHz (480x7) @ 1.50000V
This is not Orthos stable but seems to be Windows (XP not Vista) stable.

1GB (1x1GB) G.SKILL DDR2-800 (4-4-3-5) @ 1:1 @ DDR2-960 @ 6-6-6-18 @ 2.3V



Score: 9309

 

[aigomorla]

hmmm.... seems like when i increase to 3750ghz, my score doesnt go up as much as i thought it would.

:T or people here must be calling shins...

Anyhow: 10625 @ 3750ghz E6600

http://i125.photobucket.com/albums/p73/aigomorla/3-1.jpg


i dont think im gonn catch up to the quadcores :T *kickin the can*

 

[RussianSensation]

I am not liking the accuracy of this benchmark too much.

Just ran 3 back-to-back tests on E6400 @ 3400mhz.

1st run got a reading of cpu speed 3400mhz and a score of 9016
2nd run got a reading of cpu speed 5000mhz(!!!) and a score of 9195
3rd run got a reading of cpu speed 5000mhz and a score of 8897

Nothing was changed except the system benchmark ran, closed, re-ran, closed and re-ran.

 

[DrMrLordX]

Originally posted by: Arkaign

Awesome! My work PC's 1.8Ghz E4300 is exactly half that clockspeed, and it's 4800 score is almost exactly scaled to your 2x clock speed.

Conclusion : C2D architecture scales *VERY* well .. I wonder when the point of diminishing returns would hit these chips?

Sure it scales well, when you're dealing with an app that is almost completely memory/cache independent. Take a look at the way Core 2 scales in Super Pi and you'll get an entirely different story:

http://www.techspot.com/review/40-core2-e4300-vs-e6300-overclocking/page3.html

Check that out. A 3.5 ghz E6300 and E4300 scoring nearly the same as an E6700 @ stock? Yow.

 

[Accord99]

Originally posted by: DrMrLordX
Sure it scales well, when you're dealing with an app that is almost completely memory/cache independent. Take a look at the way Core 2 scales in Super Pi and you'll get an entirely different story:

http://www.techspot.com/review/40-core2-e4300-vs-e6300-overclocking/page3.html

Check that out. A 3.5 ghz E6300 and E4300 scoring nearly the same as an E6700 @ stock? Yow.

Those numbers don't look right, compared to the scores in this thread:

http://forums.anandtech.com/messageview...&STARTPAGE=3&FTVAR_FORUMVIEWTMP=Linear

 

[E6700]

9070

 

[RamIt]

Originally posted by: lyssword
3367 @2.3ghz opteron 146(2ghz stock) first run
3468 marks the second run. All hail single core king! I'm gona try to oc to 2.45

* new results (all on stock volts/stock fan)
3682 @2.45
3757 @ 2.5

Can any single cores beat this?

3932 - dell laptop single core pm 2.26

 

[aigomorla]

Originally posted by: Accord99

Originally posted by: DrMrLordX
Sure it scales well, when you're dealing with an app that is almost completely memory/cache independent. Take a look at the way Core 2 scales in Super Pi and you'll get an entirely different story:

http://www.techspot.com/review/40-core2-e4300-vs-e6300-overclocking/page3.html

Check that out. A 3.5 ghz E6300 and E4300 scoring nearly the same as an E6700 @ stock? Yow.

Those numbers don't look right, compared to the scores in this thread:

http://forums.anandtech.com/messageview...&STARTPAGE=3&FTVAR_FORUMVIEWTMP=Linear

which is WHY im starting to think, some people are calling shins!!!

But i dont know who actually is and who actually isnt. But the scores seem tad off on a lot of people.

Because hoenstly i went from 10125 @ 3636 -> 10625 @ 3750

I will believe a E6600 @ 3400 to do in the 9k ranges..

HOWEVER i refuse to believe a E6400 can do 9k sub 3.6ghz, unless this program takes no account on cache memory.

I dont know, it just seems kinda funny how some people's scores are that high, with such a moderate overclock. :T

Originally posted by: RussianSensation
I am not liking the accuracy of this benchmark too much.

Just ran 3 back-to-back tests on E6400 @ 3400mhz.

1st run got a reading of cpu speed 3400mhz and a score of 9016
2nd run got a reading of cpu speed 5000mhz(!!!) and a score of 9195
3rd run got a reading of cpu speed 5000mhz and a score of 8897

Nothing was changed except the system benchmark ran, closed, re-ran, closed and re-ran.

OMGWTFBBQ!!?!?!?!??!!?!?! 5ghz!! looks like a WR. ahahahaha j/k

yeah im not starting to like this software at all either. I think SuperPI is still a better program to see cpu computational power. 3dmark would be grafix and overall. But ive seen a crappy single core stomp a dual core just because it had a 8800GTX. so 3Dmark is a bit inacurate to test CPU power.

 

[Arkaign]

Originally posted by: Accord99

Originally posted by: DrMrLordX
Sure it scales well, when you're dealing with an app that is almost completely memory/cache independent. Take a look at the way Core 2 scales in Super Pi and you'll get an entirely different story:

http://www.techspot.com/review/40-core2-e4300-vs-e6300-overclocking/page3.html

Check that out. A 3.5 ghz E6300 and E4300 scoring nearly the same as an E6700 @ stock? Yow.

Those numbers don't look right, compared to the scores in this thread:

http://forums.anandtech.com/messageview...&STARTPAGE=3&FTVAR_FORUMVIEWTMP=Linear

No kidding, that does seem bogus! I think C2D scales spectacularly well, as did K8 when it arrived. You boost the clock by 10%, you got 10% faster cpu performance in almost everything. Pretty huge difference compared to the worn-out Athlon XP and Pentium 4 architechtures, where boosted clock speed gave ever diminishing returns.

 

[lyssword]

Originally posted by: RamIt

Originally posted by: lyssword
3367 @2.3ghz opteron 146(2ghz stock) first run
3468 marks the second run. All hail single core king! I'm gona try to oc to 2.45

* new results (all on stock volts/stock fan)
3682 @2.45
3757 @ 2.5

Can any single cores beat this?

3932 - dell laptop single core pm 2.26

Nice Pentium M is like core 2 duo with 1 core I believe. All Hail RamIt single core king !

 

[DrMrLordX]

Originally posted by: Accord99

Those numbers don't look right, compared to the scores in this thread:

http://forums.anandtech.com/messageview...&STARTPAGE=3&FTVAR_FORUMVIEWTMP=Linear

I just call it as I see it. Pi numbers taken in a closed testing environment on evenly-matched machines probably provide better comparison data than those from forum posts, but you never know. Some of those Pi scores in the Pi thread seem awfully low/inconsistent.

Originally posted by: Arkaign

No kidding, that does seem bogus! I think C2D scales spectacularly well, as did K8 when it arrived. You boost the clock by 10%, you got 10% faster cpu performance in almost everything. Pretty huge difference compared to the worn-out Athlon XP and Pentium 4 architechtures, where boosted clock speed gave ever diminishing returns.

See, that's the thing, I went looking for data on how well Core 2 chips (especially the 2 meg L2 ones) scale when overclocking. What I found was that article I linked above, and it did not show Core 2 E6300s and E4300s scaling well at all. It makes sense, too, due to the fact that they use the same memory interface that late-gen Pentium 4s and Pentium Ds used. Core 2 needs increasingly large amounts of L2 as it scales upwards in clock speed, which was the same basic situation as with P4s. Why else do you think Intel put 6 megs of L2 on Penryn? I'm sure the 4 meg L2 chips scale better than the 2 meg chips, of course, so it's not like Core 2 is completely choked off by its memory architecture. It's a weakness, but not a very big one given the current competition.

If anyone wants to post a set of Pi scores and Nuclear scores on their system at different FSB speeds (maintaining constant RAM ratio and timings if possible; use the timings from your maximum memory clock) feel free to do so. The last time I asked for scaling data in Pi I got a link to that article *P For what it's worth, I did some Pi scaling testing on an X2 awhile back and the data can be found here:

http://forums.anandtech.com/messageview...?catid=28&threadid=2016843&STARTPAGE=1

If I ever get that machine back out of its box before shipping it off somewhere, maybe I could do some Nuclear 1.50 testing as well. But for now, it stays in the box.

 

[swtethan]

8559

 

[adairusmc]

I just got my new C2D e6600 installed this evening, and got around 5780. I think it is because I am using the same windows install as I had for the AthlonX2 system (I am waiting for Vista x64 to arrive, to replace the 32bit version I have now). I will run it again after I get this thing reformatted with vista x64.

 

[Accord99]

Originally posted by: DrMrLordX
I just call it as I see it. Pi numbers taken in a closed testing environment on evenly-matched machines probably provide better comparison data than those from forum posts, but you never know. Some of those Pi scores in the Pi thread seem awfully low/inconsistent.

The thread matches what Anandtech gets:

http://www.anandtech.com/mb/showdoc.aspx?i=2914&p=2

At stock, an E6300 needs ~28s for 1M digits, which is comparable to the techspot review. However, once overclocked to 3.5GHz with DDR2-1000, Anandtech gets a time in the 16s range.

 

[krotchy]

Ok, 2K workstation below, Dual Xeon 5355's

ALU Score: 4806
FPU Score: 5155
MT Score 8T:24385

Total Score: 17563

 

[Icehawk]

With a ton of stuff open (not sure if that matters) my E6600 @ 3.25/900mhz hit 9092 which is higher than a bunch of others at similar clockspeed - weird.

Ok ran with a cleanish boot and hit 9242. My e-penis is hard.

 

[DrMrLordX]

Originally posted by: Accord99

http://www.anandtech.com/mb/showdoc.aspx?i=2914&p=2

At stock, an E6300 needs ~28s for 1M digits, which is comparable to the techspot review. However, once overclocked to 3.5GHz with DDR2-1000, Anandtech gets a time in the 16s range.

That's odd . . . I wonder what differences there were between the Anandtech and Techspot benchmarking machines?

 

[GuitarDaddy]

I agree this bench shows inconsistencies. I just ran it again with the exact same settings as my first post E6400 at 3.280ghz and this time I get 9173, when I got 8782 before.

 

[Arkaign]

Originally posted by: Accord99

Originally posted by: DrMrLordX
I just call it as I see it. Pi numbers taken in a closed testing environment on evenly-matched machines probably provide better comparison data than those from forum posts, but you never know. Some of those Pi scores in the Pi thread seem awfully low/inconsistent.

The thread matches what Anandtech gets:

http://www.anandtech.com/mb/showdoc.aspx?i=2914&p=2

At stock, an E6300 needs ~28s for 1M digits, which is comparable to the techspot review. However, once overclocked to 3.5GHz with DDR2-1000, Anandtech gets a time in the 16s range.

Yeah, obviously I agree

It looks like the 4300/6300/6400 scale VERY well up to the current top frequencies (4Ghz or so). Every architecture runs into a scaling problem at some point, but I don't see any such evidence of anything like that on the C2D 2M and 4M procs at this point. Basically, give 10% more clock speed, get close to 10% more raw cpu performance.

 

[Icehawk]

 

[Icehawk]

For giggles, here is my work PC - Xeon 3.4ghz, 2gb Ram at stock everything:

3,873

Thanks for getting me a Xeon and not a C2D more money for less power.

 

[bruceyg]

My Athlon 3600+ @2.9Ghz, 1gb ddr2 667 RAM

scores 5760, happy with it, it's only a $75 processor

 

[DukeN]

E6420 @ 2.88
4 Gig Ballistix @ 4-4-4-12 950mhz

8571 Marks

 

[montypythizzle]

Can't DL it, or see the SS.