3DMark '13 score thread

[malfunction]

Latest score from my main rig, using a clean install of the latest (non-beta) nvidia drivers:

http://www.3dmark.com/3dm/256128

(GPU @ 1050MHz, mem @ 2350MHz)


Run from another i7 9xx machine with an old Radeon HD3850 (which can't run fire strike):

http://www.3dmark.com/3dm/255675
http://www.3dmark.com/3dm/255745 (with a slight overclock)

[scaryjim]

you're just competing against yourself now jimbob

And compete I will

Currently installing on the main rig, drivers and downloading, and I've got a new BIOS sitting on a USB stick as it won't boot with my new memory (soon sort that out though, hopefully ). Probably won't get much more done than Cloud Gate on the HD4200 tonight, but Friday I should have some time to mess with the 3470, look at Hybrid Crossfire, and perhaps give you all a nice surprise too

EDIT:
Hmmm, 3470 is fighting me. Can use a different low profile graphics card (any guesses as to what that is, anyone? ) and get output from it, but when I put the 3470 in, no output. It's worked in the past on a different motherboard no problems. Any thoughts? It's a Dell, or possibly HP, OEM 3470 with dual displayport, on an ASUS M4A785TD-M Evo...

In the meantime, here's what a 300MHz boost in core speed can do for HD4200 integrated:

Stock settings (HD4200 @ 500MHz, PII 905e @ 2.5GHz): http://www.3dmark.com/3dm/261925
Overclocked (HD4200 @ 800MHz, PII 905e @ 2.5GHz): http://www.3dmark.com/3dm/262017

300MHz overclock gets me above stevie's HD3300 Cloud Gate score - result

Now if only I could get this 3470 working...

[scaryjim]

OK, time to complete the DX10 comparison

First up, the issue with the 3470 is that it apparently doesn't have any analogue output conversion ability on the card. My passive DP - VGA adapter (thanks Biscuit ) works fine in two other DP devices, but not my 3470. No worries, I've just moved the rig downstairs to a DVI monitor, and it's worked fine

So, the HD3470 on its own: http://www.3dmark.com/3dm/264546 - ignore the Graphics section of the result, it's got confused by there being 2 GPUs in the system: the gfx is actually an HD3470 256MB @ 750MHz

Then it was time to try Hybrid Crossfire My first attempt crashed the PC - oh noes! So I dialled the IGP clock speed back to default (500MHz) and ran that: http://www.3dmark.com/3dm/264414 - Again, ignore the graphics: this is HD3470 + HD4200, crossfire, @ 500MHz

And, having discovered that my 3470 apparently runs at 750MHz, not 800 as I'd previously assumed, I tried the Hybrid agaion with the IGP at 750Mhz: http://www.3dmark.com/3dm/264478 - This is 3470+4200 @ 750MHz.

You know, I might write up a user review of all those results. Main takeaways are: HD4200 @ 800MHz with sideport is *faster* than a discrete 3470, and that Hybrid Crossfire really works

[Biscuit]

We need a table update

[kalniel]

We need a table update

Steady on. He did update quite recently - it's only a volunteer effort after all

[Biscuit]

I expect tri-weekly updates!

[Ulti]

Woo, got my new CPU cooler installed a few days back and upgraded GPU drivers. Not sure if my CPU was being throttled before but it it was then hurray for no more throttling!

From http://www.3dmark.com/3dm/81246 to http://www.3dmark.com/3dm/266177.

15.5% increase in Fire Strike to 3953. Not bad!

[KeyboardDemon]

15.5% increase in Fire Strike to 3953. Not bad!

Congrats, that's a good result. Though not sure how much of the increase was down to new drivers, would be interesting to see what you get if you could roll back to previous drivers and test again.

[Biscuit]

Congrats, that's a good result. Though not sure how much of the increase was down to new drivers, would be interesting to see what you get if you could roll back to previous drivers and test again.

The new Nvidia drivers made a pretty juicy difference for me. In single GPU mode i saw a boost of about 500 and in SLI it actually worked

[rickyjb]

score 11472 http://www.3dmark.com/3dm/274879

fire strike extreme score 6010 http://www.3dmark.com/3dm/275191 (that combined test dragged along at about 10 fps)

cpu: Intel Core i7-980 @ 4,173 MHz
gpu: 2 x gtx 680 , boost core clocks 1188

[Ulti]

Congrats, that's a good result. Though not sure how much of the increase was down to new drivers, would be interesting to see what you get if you could roll back to previous drivers and test again.

Did so for curiosity's sake:

Throttling with whatever W8 installed for drivers ("9.2.0"): http://www.3dmark.com/3dm/81246
No throttling with whatever W8 installed for drivers ("9.2.0"): http://www.3dmark.com/3dm/275458
No throttling with latest stable drivers ("9.12.0"): http://www.3dmark.com/3dm/266177

So that's a 7.3% increase in score from upgrading the drivers alone. Gonna try the beta drivers but I doubt it'll be any better than the latest "stable" drivers.

Will edit this post with beta drivers later in a bit.

EDIT: Latest beta drivers show no improvement: http://www.3dmark.com/3dm/275670

[scaryjim]

Well, I promised a nice surprise for you all, so here it is:

http://www.3dmark.com/3dm/272457

Obviously the Cloud Gate is the interesting result for me, to do the comparisons with, but I'm pretty happy with 1985 in Fire Strike, from a slimline matx build that draws less power at full load than my old Q6600/880GTX drew at idle

[shootomanUK]

3dmark.com/3dm/288580

i7 920 @ 3.8ghz
gtx480 sli @ stock




(sorry cant post link to 3DMARK not enough posts lol)

[Bagnaj97]

The post got too long so I was going to move the Cloud Gate table to a new thread, except I seem to have deleted it when I meant to cut and by the time I realised it was too late!

As for the 7870/7870 LE cards, it doesn't look like 3dmark distinguishes. I've been getting the card info from 3dmark and changing the clocks where people have said they're different. If people let me know whether their card is a 7870 or 7870 LE I'll update the table.

[Phage]

Woo - Still top of the single GPUs.

[CAT-THE-FIFTH]

The table has a mistake. It is an A6-3670K not an A4-3670.