AMD Benchmarks Zacate APU, 2x Faster GPU Performance than Core i5

[richierich]

Clearly so, jimbo. The laptop must have not had the power saving GPU setting set to performance as someone else speculated.

Guy said it was set to performance mode?

[Manicdan]

maybe it overheated playing CoH

[god_43]

well the ie9 beta is released tomorrow, i think that the preview is a little screwy.

[AKM]

Direct2D acceleration doesn't look to be enabled at all for the Intel system.

[Hornet331]

well the ie9 beta is released tomorrow, i think that the preview is a little screwy.

Which doesn't matter since everyone is using the same software. And we even saw a GMA 4500 scoring the same as a GMA HD or HD5870.

The presentation was bonked.

[Particle]

Guy said it was set to performance mode?

If he isn't mistaken, then there aren't a lot of other explanations that I can think of. The i3 can't perform that badly unless something is horribly wrong.

[informal]

I think that maybe the intel GPU driver doesn't properly support IE9 hw acceleration.The rest of the presentation should be representative of i5's GPU performance.

[god_43]

Which doesn't matter since everyone is using the same software. And we even saw a GMA 4500 scoring the same as a GMA HD or HD5870.

The presentation was bonked.

yeah your right, something is messed up? i mean they could have picked better tests i think? apparently this is in direct x 9 right? why, i mean cant we move past it?

[kaktus1907]

GTX 460

[Manicdan]

anyone notice that most numbers are very similar, almost looks like a fps cap to me

[Chumbucket843]

my OCed i7 920 + GTX260 scores ~650. cpu utilization is about 3-4% and gpu utilization is ~35% while my notebook with an i7 and a 5730 scores 580. perhaps there is limited parallelism in this bench hence low end and high end hardware perform similarly.

[Hornet331]

anyone notice that most numbers are very similar, almost looks like a fps cap to me

Aero is vsynced.

[informal]

perhaps there is limited parallelism in this bench hence low end and high end hardware perform similarly.

That was my guess too.Something is limiting the scores across the platforms.

[Manicdan]

looking like a test that just says, good enough, or something be messed up

[jimbo75]

Guys you are aware this is the 18w i5 that was being tested, yes?

http://www.youtube.com/watch?v=XH5A4D9qoDQ forward to 1:47 and it looks like it's the UM being tested.

[dreamdx]

Aero is vsynced.

Doesnt matter in this case.
I get the same results @ 75 vs 60hz. Windows Classic doesnt change anything, neither does killing WDM.exe.
I tried playing a 720p MKV @ DXVA via EVR + AERO enabled at the same time while running the test and no change. Always 606 points in Hallucinogenic mode.

BTW, HD4290 256MB Shared. Dont ask ...

[informal]

Guys you are aware this is the 18w i5 that was being tested, yes?

There is no 18W model that runs @ 2.4Ghz,like the one in the demo.

[jimbo75]

There is no 18W model that runs @ 2.4Ghz,like the one in the demo.

Can you point to the part in the video where he says its 2.4ghz informal? It's not totally clear but if you listen to about 1:45 onwards it seems that they are talking about the i5 @ 18w - the first guy asks and he is pointing the camera at the i5, then the AMD seems to point his hand at it and say "this is 18w".

edit - just wondering if the 2.4ghz thing was suggested in the press or if it was mentioned in any of the demo's? Can anybody find anything in any of the demo's that suggests the i5 is running at 2.4ghz?

[informal]

I thought that I saw the 2.4Ghz figure in one of the videos,but I can't find it now... So after all,it may be the same 18W i5 system,in which the GPU runs at very low clocks and CPU is in the 1.2-2Ghz(Turbo) range.It fits with the scores at least... And it really is logical to compare 18W i5 vs 18W Zacate,same power band etc.

[enumae]

Processor used - Intel® Core™ i5-520M

Video in which he specifies (1:45'ish)

[jimbo75]

Processor used - Intel® Core™ i5-520M

Video in which he specifies (1:45'ish)

Hmm interesting, though that could be an easy mistake to make.

Is there any evidence of 2.4ghz being mentioned? That would pretty much nail my theory stone dead.

[informal]

Thanks enumae,I've missed it. M520 has a 500Mhz GPU and the whole chip is 35W,not really a competitor to Zacate. The 18W i5 is more appropriate for comparison IMO,but that one has really low CPU and GPU clocks.


edit: maybe the guy actually meant Intel® Core™ i5-520UM ,it's easy mistake to make,one letter. This one fits perfectly,it's 18W and has low clocks.
http://ark.intel.com/Product.aspx?id=47554

Graphics Base Frequency 166 MHz
Graphics Max Dynamic Frequency 500 MHz

Clock Speed 1.066 GHz
Max Turbo Frequency 1.866 GHz

Max TDP 18 W

[jimbo75]

Thanks enumae,I've missed it. M520 has a 500Mhz GPU and the whole chip is 35W,not really a competitor to Zacate. The 18W i5 is more appropriate for comparison IMO,but that one has really low CPU and GPU clocks.


edit: maybe the guy actually meant Intel® Core™ i5-520UM ,it's easy mistake to make,one letter. This one fits perfectly,it's 18W and has low clocks.
http://ark.intel.com/Product.aspx?id=47554

Lol, this just gets more confusing

I think it is the UM. This 10% score in the browser thing in 2d might be indicative of the seriously low 2d clock of 166mhz, whereas the game City of Heroes would at least turbo the graphics up to 500mhz, meaning it's about half as good as this Zacate chip (and that is also what we are seeing).

I'm ruling nothing out however.

[Hornet331]

zacate.jpg

http://www.youtube.com/watch?v=-lLNHTu4oLk


Still doesn't explains the botched results.

[QuadDamage]

no.. the question is what will bulldozer do to sandy bridge ...

Just a feeling but prolly nothing.