First (real) GMA X3000/G965 Benchmarks

[neyn]

Look what I found..

I believe only these linux benchmarks mean anything right now, since the public win drivers haven't been updated. Comparing these numbers (reference scores here) shows G965 about 20-40% slower than the $400 cards.. not bad for an integrated chip.

 

[Skrying]

Look what I found..

I believe only these linux benchmarks mean anything right now, since the public win drivers haven't been updated. Comparing these numbers (reference scores here) shows G965 about 20-40% slower than the $400 cards.. not bad for an integrated chip.

Yes, in a game that terribly CPU limited and also when the benchmarks are run at only 1024x768 (again, CPU limited) and at generally low detail settings (hey look, CPU limited again!).

 

[Geeforcer]

CPU limited? No-no-no, the correct term is "real-world, AEIABOFL*".






*Anything Else Is A Bunch of Filty Lies.

 

[Jakob]

More G965 news at digitimes:

Asustek G965-based motherboards with Shader Model 3.0 support expected later

It mentions that G965 boards have been shipping since August (sans SM3 drivers). So it is curious why there are still no published benchmarks.

Newegg even lists 5 G965 mobos in stock:

http://www.newegg.com/Product/Produ...ion=G965&Submit=ENE&N=0&Ntk=all&Go.x=0&Go.y=0

Comon B3D, here's your chance to break the news about X3000 perf.

 

[Jakob]

I found these month-old G965 benchmark scores on a chinese site:

http://www.pconline.com.cn/market/sh/shoppingguide/changshang/0608/844892.html

It is slower than 945G across the board. Immature drivers, perhaps, but this doesn't bode too well for intel's first unified shader effort.

I still would like to see a more comprehensive test with the latest drivers.

 

[Jakob]

found some more recent (8/29) benchmark results here:

http://www.pconline.com.cn/diy/evalue/evalue/main/0608/856535_4.html

Overall, looks about a wash for real world gameplay compared to 945G. ie still single digit fps in games like d3 and fear. At least it does run SM3 tests in 3dmark06.

 

[Zaphod]

Immature drivers, perhaps, but this doesn't bode too well for intel's first unified shader effort.

Digitimes reported a while a go that Intel was going for a respin to fix graphics problems, and that this revision wouldn't be widely shipping to ODM partners before late August. That would certainly explain the lackluster availability, but whether performance will be improved from the few benchmarks that are out there is, perhaps, questionable. I wouldn't rule it out, though.

 

[neyn]

Digitimes reported a while a go that Intel was going for a respin to fix graphics problems, and that this revision wouldn't be widely shipping to ODM partners before late August. That would certainly explain the lackluster availability, but whether performance will be improved from the few benchmarks that are out there is, perhaps, questionable. I wouldn't rule it out, though.

DigiTimes actually said: "sources at motherboard makers stated that Intel's driver sets released in August did not enable Shader Model 3.0 support and the lack of the functionality was a factor for postponed shipments from some motherboard makers"; it's not the hardware but the software.
The chinese benchmarks above are all on the wrong drivers, so what we really need is for someone to buy a G965 mb and try SPECviewperf or another cross-platform test.

 

[Zaphod]

DigiTimes actually said: "sources at motherboard makers stated that Intel's driver sets released in August did not enable Shader Model 3.0 support and the lack of the functionality was a factor for postponed shipments from some motherboard makers"; it's not the hardware but the software.

Not that article, an older one which is also referenced a couple of paragraphs below the text you quoted.

motherboard makers expected the G965 to be ready for volume shipments in the end of July, at the same day with the launch of desktop Core 2 Duo (Conroe) processors. However, a defect in the integrated graphics subsystem was found, and shipments were postponed, the makers indicated.

Edit: But apparently Intel says the above isn't true, and also says the drivers are A'OK. For their sake, I hope they're lying.

 

[RejZoR]

CPU limited? Lol. You can be CPU limited on 7950GX2, not on this piece of crap.

 

[swaaye]

Powered by Gigabyte Triton III System

u really running a powerhouse 430VX?

 

[Sc4freak]

CPU limited? Lol. You can be CPU limited on 7950GX2, not on this piece of crap.

Exactly. That's why the integrated solution looked so fast compared to the top end cards.

 

[Skrying]

Exactly. That's why the integrated solution looked so fast compared to the top end cards.

Thank you for stating my point.

 

[RejZoR]

u really running a powerhouse 430VX?

Hehe AMD Athlon XP on Intel chipset. Don't think it'll ever happen
It used to have pretty much all components from Gigabyte and it was my 3rd generation PC. I've already sold the Gigabyte graphic card, Gigabyte mobo is still there (Triton series, thats why name Triton ). I guess i was just to lazy to change the sig

 

[TraumaPill]

I just got the G695SS motherboard (which includes the X3000 IGP) and I'd be glad to run the benchmarks, since that is what I was looking for when I found this forum, if you guys will hook me up with links and instructions. It may be up to a week before I can get around to it but I will certainly do it when I have time.

I am currently running a Core2Duo 1.86 Mhz processor and 512 MB DDR2.

 

[DavidC]

The best thing is to have a user with GMA950 bench the same UT on Linux. Then we will know whether its CPU depency or graphics advancement.

I just got the G695SS motherboard (which includes the X3000 IGP) and I'd be glad to run the benchmarks, since that is what I was looking for when I found this forum, if you guys will hook me up with links and instructions. It may be up to a week before I can get around to it but I will certainly do it when I have time.

I am currently running a Core2Duo 1.86 Mhz processor and 512 MB DDR2.

I've got a request. Run any 3dmark program(2001, 2003, 2005 etc.), and post the details of the benchmark, particularly the fillrate. Intel's graphics go pretty close to maximum fillrate so we can derive the max fillrate by measuring it.

If you can do 1024x768 Quake 3 benchmark, that would be cool too. :smile:

Check the driver versions and benchmark them on both 14.21 and 14.24 drivers, we'll see if that does anything.

 

[TraumaPill]

Hmm. I can't seem to find the version 14.24 drivers despite the fact that they've supposedly been released. All I found was 14.21 on Intel's site.

 

[ronin-2]

Driver link for the 14.24 is here, saw it on Station Drivers Intel page.

 

[TraumaPill]

I can't seem to connect to that site to download those drivers.
However I'll go ahead and try to test the 14.21 drivers, I'm currently at 14.10 (I haven't really had a chance to test the card because I've been too busy to play games).

 

[TraumaPill]

Ok, so I acquired 3DMark06 and tried to run it, but it keeps crashing when I try to run the graphics test - it goes to load them, and after it finishes the screen goes black, then after a few seconds the process is dead and it closes. I had ended all nonessential processes and disabled my network connections...it could be the copy I got. I don't know. Any ideas?