Wednesday, April 14th 2010

ASUS Radeon HD 5870 Sets GPU Speed Record of 1525 MHz (core)

In an overclocking feat by XtremeSystems community members Elmor and Kinc, an ASUS EAH5870 MATRIX Radeon HD 5870 graphics card was able to achieve a mind-boggling core speed of 1525 MHz (core) from its default speed of 894 MHz, and AMD reference speed of 850 MHz. This translates into 70% over default speeds, and 79% over reference speeds. The not-so-relevant memory overclock was 1300 MHz (5.2 GHz effective), an 8.3% overclock. The card was cooled using liquid-nitrogen evaporators.

This wasn't a "hit-and-run" feat, either. The stability of the GPU at 1500 MHz was proven with a run of 3DMark Vantage, where it scored P29342 points at the performance preset. The rest of the bench comprised of an Intel Core i7 980X Extreme Edition six-core processor running at 4492.9 MHz, at 1.52V, also cooled by liquid nitrogen, ASUS Rampage III Extreme motherboard, 6 GB of Corsair Dominator GTX DDR3 memory, and Antec 1200W power supply. The validation for the 3DMark Vantage run where it scored P30542 can be found here.
Source: XtremeSystems Forums
Add your own comment

45 Comments on ASUS Radeon HD 5870 Sets GPU Speed Record of 1525 MHz (core)

#26
Hayder_Master
H82LUZ73Holy god Lets see if Fermi can match that.:toast:
Fermi :laugh: , fermi need LN2 to hit 1ghz maybe
ObSo-1337wow... i wonder what voltage it was using :eek:
me too i think they going over 1.3v
Posted on Reply
#29
Wile E
Power User
jasper1605it's just a curiosity. It seems like there's always more emphasis placed on core speed rather than mem speed so I'm just wondering why.
Core speed make more of a difference on the score.
leonard_222003But it can't be validated because it uses the cheating drivers that increases Cpu score.
What are you talking about? If you mean Physx, it was off for that run. Not to mention, HT was shut off, and he was only clocked to 4070 on the cpu.
Posted on Reply
#30
leonard_222003
Wile ECore speed make more of a difference on the score.



What are you talking about? If you mean Physx, it was off for that run. Not to mention, HT was shut off, and he was only clocked to 4070 on the cpu.
ARE YOU BLIND MAN ? LOOK AT THE SCORE OF A 4.4 GHZ CORE I7 ON THE ATI CARD AND THEN AT A 4 GHZ CORE I7 ON THE FERMI CARD
www.techpowerup.com/img/10-04-14/52a.png
j.imagehost.org/view/0899/31704

THERE IS A 12910 difference and the Ati card has a faster clocked CPU !!!
:D:D:D Blind fa..... in denial of course.
Posted on Reply
#31
WhiteLotus
Jesus. That is some hardcore overclocking. In reality that's like looking in the type of scores that one might get in the near future. All they need is a better way of dissipating heat and we're set.
Posted on Reply
#32
Wile E
Power User
leonard_222003ARE YOU BLIND MAN ? LOOK AT THE SCORE OF A 4.4 GHZ CORE I7 ON THE ATI CARD AND THEN AT A 4 GHZ CORE I7 ON THE FERMI CARD
www.techpowerup.com/img/10-04-14/52a.png
j.imagehost.org/view/0899/31704

THERE IS A 12910 difference and the Ati card has a faster clocked CPU !!!
:D:D:D Blind fa..... in denial of course.
Look at CPU test number 2. That IS NOT a physx score, period. If Physx was on, that test would be in triple digit frame rates. Even my 8800GT gets 130fps in that test. service.futuremark.com/compare?3dmv=1797491

Here is one on a GTX480 with Physx on: service.futuremark.com/compare?3dmv=2090491 Notice it gets almost 275fps in CPU Test 2.

Not to mention, look at GPU-Z, the Physx box is unchecked on the Fermi score, meaning Physx is disabled.

Do some research before making baseless claims.
Posted on Reply
#33
aj28
Wile ELook at CPU test number 2. That IS NOT a physx score, period. If Physx was on, that test would be in triple digit frame rates. Even my 8800GT gets 130fps in that test. service.futuremark.com/compare?3dmv=1797491

Here is one on a GTX480 with Physx on: service.futuremark.com/compare?3dmv=2090491 Notice it gets almost 275fps in CPU Test 2.

Not to mention, look at GPU-Z, the Physx box is unchecked on the Fermi score, meaning Physx is disabled.

Do some research before making baseless claims.
QFT. ATi run has HT on.
Posted on Reply
#34
kid41212003
Both CPU 1 (1 is not physX) and 2 tests are higher, even though it has lower clock speed.

I say the guy with higher clock and lower CPU scores has a fucked up CPU.
Posted on Reply
#35
Kreij
Senior Monkey Moderator
Yeah, but will it run DOS??? Oh wait ...

That is quite an impressive OC.
Posted on Reply
#36
Wile E
Power User
kid41212003Both CPU 1 (1 is not physX) and 2 tests are higher, even though it has lower clock speed.

I say the guy with higher clock and lower CPU scores has a fucked up CPU.
Or Vantage may not like HT on a 6 core, or he didn't tweak his install as far. Could be a multitude of things. Maybe they even just pushed the cpu a little too far on that voltage, and might get a better score with a slightly lower clock. Who knows?

Either way, both OCs are damn impressive.
Posted on Reply
#37
qubit
Overclocked quantum bit
I don't hold a 3D Mark score in much regard when comparing different architectures eg ATI v nvidia or even different generations from the same company, as the results don't tend to match real-world gameplay all that well.

So the fact that the GTX 480 beats that tortured 5870 don't mean jack to me - I want to see real game scores or it's meaningless.

The only time it really means something in my book is when you're comparing the same chip under different clocking conditions or the same card from different manufactures ie two GTX 480s or something. That result by Shamino is only comparable to another 5870.
Posted on Reply
#38
btarunr
Editor & Senior Moderator
Wile EWhat are you talking about? If you mean Physx, it was off for that run. Not to mention, HT was shut off, and he was only clocked to 4070 on the cpu.
That's a dual-processor setup. Two processors on Classified SR-2. Look at the "processor" dropdown. So there are 12 cores even if HTT was disabled (still faster than a single CPU on HTT).
Posted on Reply
#39
Kitkat
yogurt_21interesting that this is basically what a 5970 would perform at stock with that same cpu/memory overclock. at least based on the toms review of the gtx480
www.tomshardware.com/reviews/geforce-gtx-480,2585-6.html
www.tomshardware.com/reviews/geforce-gtx-480,2585-7.html

nice oc but interesting to see 1600 sp's at 1500MHZ vs 3200sp's at 725mhz = the same performance
yeah it has 2 gigs of mem u want a comparison of 5970 stock with Aries. (4gb) (oced by the same person ofcourse ;) )
Posted on Reply
#40
Wile E
Power User
btarunrThat's a dual-processor setup. Two processors on Classified SR-2. Look at the "processor" dropdown. So there are 12 cores even if HTT was disabled (still faster than a single CPU on HTT).
Thanks. I missed that detail. Gives even more merit to my point.
Posted on Reply
#42
Lu(ky
qubitDespite this massively impressive overclock, a GTX 480 will be faster and better. This is because nvidia always make the fastest and best cards, bar none.

I'd read the rest from this nvidia PR leaflet, but the dog ate it. :laugh:
All you need is something to cool that GTX 480. I suggest this picture below? Or do I need to go bigger.. :D Also nice oc on the card, but the CPU at 1.5v and only at 4.5 seems low to me. Or he has a bad ocing chip.

Posted on Reply
#43
WSP
I do feel pity for the polar bear

Posted on Reply
#44
fatguy1992
Guys the reason why his CPU score is so high is because he has 2x 6core guftown's. He turned HT off as vantage can only use 12 core/threads. Its not phyx or a wrong score. Sorry if someone else already posted this, i had a quick scan through comments.
Posted on Reply
#45
leonard_222003
Well i missed that too 2x 6core guftown's, i was under the impression it has physx enabled or something , eighter way it doesn't matter , the GTX480 has an advantage and the score difference would be so little if they were on the same platform , sure the GTX480 would win but by a hair.
Posted on Reply
Add your own comment
Apr 25th, 2024 00:42 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts