3DMark (2013) Scores Thread

W0LFMAN

Boards.ie 3DMark (2013) Leader Board.




ICESTORM

Rank |Member Name|Score - Rig Specs
1.|SpannerMonkey|180,798 - (I7 3770k (CPU Clock 4.50Ghz), 16g Ram, GTX 690)
2.|Bloodbath|177,300 - (I5 3570k (CPU Clock 4.50Ghz), 8g Ram, Radeon HD 7970)
3.|Serephucus|175,793 - (I5 3570k (CPU Clock 4.60Ghz), 16g Ram, GTX 670)
4.|Deano12345|174,417 - (I7 2600k (CPU Clock 4.80Ghz), 32g Ram, GTX 680)
5.|Sickboy|140,748 - (I7 2600 (CPU Clock 4.10Ghz), 16g Ram, GTX 670)
6.|W0LFMAN|128,509 - (I7 990x (CPU Clock 3.60Ghz), 24g Ram, GTX 480)
7.|ED_E|106,327 - (Quad Core Q6600 (CPU Clock 3.50Ghz), 8g Ram, Radeon HD 7870 )
8.|Marco_Polo|104,438 - (AMD Phenom II X4 965 (CPU Clock 3.70Ghz), 8g Ram, Radeon HD 7850 )
9.||
10.||
11.||
12.||
13.||
14.||
15.||


CLOUDGATE

Rank |Member Name|Score - Rig Specs
1.|SpannerMonkey|27,870 - (I7 3770k (CPU Clock 4.50Ghz), 16g Ram, GTX 690)
2.|Deano12345|24,989 - (I7 2600k (CPU Clock 4.80Ghz), 32g Ram, GTX 680)
3.|Sickboy|21,266 - (I7 2600 (CPU Clock 4.10Ghz), 16g Ram, GTX 670)
4.|W0LFMAN|19,645 - (I7 990x (CPU Clock 3.60Ghz), 24g Ram, GTX 480)
5.|Bloodbath|19,568 - (I5 3570k (CPU Clock 4.50Ghz), 8g Ram, Radeon HD 7970)
6.|Serephucus|19,427 - (I5 3570k (CPU Clock 4.60Ghz), 16g Ram, GTX 670)
7.|ED_E|12,734 - (Quad Core Q6600 (CPU Clock 3.50Ghz), 8g Ram, Radeon HD 7870 )
8.|Marco_Polo|12,008 - (AMD Phenom II X4 965 (CPU Clock 3.70Ghz), 8g Ram, Radeon HD 7850 )
9.||
10.||
11.||
12.||
13.||
14.||
15.||



FIRESTRIKE

Rank |Member Name|Score - Rig Specs
1.|SpannerMonkey|9,693 - (I7 3770k (CPU Clock 4.50Ghz), 16g Ram, GTX 690)
2.|Bloodbath|7,505 - (I5 3570k (CPU Clock 4.50Ghz), 8g Ram, Radeon HD 7970)
3.|Deano12345|6,758 - (I7 2600k (CPU Clock 4.80Ghz), 32g Ram, GTX 680)
4.|Serephucus|6,300 - (I5 3570k (CPU Clock 4.60Ghz), 16g Ram, GTX 670
5.|Sickboy|5,830 - (I7 2600 (CPU Clock 4.10Ghz), 16g Ram, GTX 670)
6.|Marco_Polo|4,557 - (AMD Phenom II X4 965 (CPU Clock 3.70Ghz), 8g Ram, Radeon HD 7850 )
7.|ED_E|4,515 - (Quad Core Q6600 (CPU Clock 3.50Ghz), 8g Ram, Radeon HD 7870 )
8.|W0LFMAN|2,960 - (I7 990x (CPU Clock 3.60Ghz), 24g Ram, GTX 480)
9.||
10.||
11.||
12.||
13.||
14.||
15.||


(* thanks to Serephucus for table idea )

Rules to enter:-

Information required is outlined below. Anything missing = invalid post and leaderboard will not be updated.

Please provide a screenshot of:
Two (2) instances of CPU-Z(Memory and CPU Tabs)
GPU-Z,
ORB site with your score showing
An open copy of Notepad with the date and your BOARDS.IE username


In the post include:
Driver version
CPU/Clockspeed/GPU Model/Number of GPU Cores/GPU Core Clock/GPU Memory Clock

Downloads link:

Free Version of 3DMark.
http://www.guru3d.com/files_details/3dmark_download.html

GPU-Z Download Link
http://www.techpowerup.com/gpuz/

CPU-Z Download Link.
http://www.cpuid.com/softwares/cpu-z.html


RULES:

NO AMD vs nVidia Fighting
NO unsubstantiated claims without references/links
NO Spaming Please

W0LFMAN

ED E

I'll bite:

http://www.3dmark.com/3dm/73428

ICE: 106327
CLOUD: 12734
FIRE: 4515



XFX 7870 Beta 13.1 Stock clocks
Q6600 @ 3.5
2x4GB @650Mhz
128GB 830


If this thread actually takes off I'll be way down the chart.

Serephucus

Wow that took a long time to run.

Ice: 176,345
Cloud: 19,386
Fire: 6,052

marco_polo

Card: HIS 7850 OC to 1,100MHz(core)/1,350MHz(memory).

Screenshot: https://us.v-cdn.net/6034073/uploads/attachments/79932/239764.jpg

Scores:
Ice Storm: 104438
Cloud Gate: 12008
Fire Strike: 4557

EDIT: Just noticed an annoying GPU-z screenshot popup is kind of blocking it but drivers version was 13.1

Deano12345

So it seems 3DMark did a run then didnt upload my details properly even though the scores validated. I'll run a proper one tomorrow but this will do as a placeholder :

Card : GTX680 Lightning @ 1202/6008 (Stock) on 313.96 drivers

Scores :

Ice Storm : 174417
Cloud Gate : 24989
Fire Strike : 6758

Screenshot : https://us.v-cdn.net/6034073/uploads/attachments/268008/239771.jpg

Since 3DMark didnt upload any details, I'll fill them in here I guess !

CPU is a 2600K, 4.8GHz on a UP5-TH (with F12G beta bios).
4x8GB of RAM @ 1866MHz
Corsair H100 cooling with P/P SP120 fans

BloodBath

Card - Sapphire 7970 dual-x.

Clocks - 1200/6200. Cata 13.2 beta driver.

Virtu MVP was enabled with hyperformance. I'll have to try another run with it disabled to see how it compares.

Scores;

Ice Storm : 177300
Cloud Gate : 19568
Fire Strike : 7505

Screenshot: https://us.v-cdn.net/6034073/uploads/attachments/21470/239774.jpg

Deano12345

@BloodBath : How's your vCore so low man ? Im jealous here

BloodBath

It drops down while idling. The cpu speed is supposed to drop down as well but it's not for some reason.

I'll have to check that. It's 1.2 under load. I can get to 4.6@ 1.3 but it's not worth it.

Looks like the 680gtx is about 50% better at the physics tests. The ati card makes up the ground in other areas. I don't understand why ati cards are still so far behind on the physics side. This benchmark doesn't use phys-x for the physics parts does it? Or is it using cpu based physics calculations? Your hyper threading might explain your performance there.

Deano12345

BloodBath wrote: »

It drops down while idling. The cpu speed is supposed to drop down as well but it's not for some reason.

I'll have to check that. It's 1.2 under load. I can get to 4.6@ 1.3 but it's not worth it.

Ah right, mines the opposite, the clocks drop down on idle but not the voltage. I've only gotten mine stable at 4.8 with this new board, the old one just couldn't do any more than 4.5 for some reason

BloodBath

Deano12345 wrote: »

Ah right, mines the opposite, the clocks drop down on idle but not the voltage. I've only gotten mine stable at 4.8 with this new board, the old one just couldn't do any more than 4.5 for some reason

I've edited that last post since

It's back downclocking to 1600mhz when idle. I think cpu-z is just bugged.

Deano12345

BloodBath wrote: »

I've edited that last post since

It's back downclocking to 1600mhz when idle. I think cpu-z is just bugged.

Yeah I think clock for clock the 680 and 7970 are pretty even. I'll see what I can do about overclocking the 680 I guess Interesting about the Physics part though, I'll do a run with HT off tomorrow and see what happens

marco_polo

BloodBath wrote: »

It drops down while idling. The cpu speed is supposed to drop down as well but it's not for some reason.

I'll have to check that. It's 1.2 under load. I can get to 4.6@ 1.3 but it's not worth it.

Looks like the 680gtx is about 50% better at the physics tests. The ati card makes up the ground in other areas. I don't understand why ati cards are still so far behind on the physics side. This benchmark doesn't use phys-x for the physics parts does it? Or is it using cpu based physics calculations? Your hyper threading might explain your performance there.

If this test ( I wish not mine , from overclock.net) with the same card clocked a shade higher and paired with an i7-3970X is anything to go by, I'd say the physics test is a little bit CPU bound (I also saw another HD7970 paired with an i7-3770K get about 12500ish on the physics too, which is in the same ballpark as Deano12345 i7-2600k).

Deano12345

I'm kinda surprised on the increase in Fire Strike with a 350MHz bump to the GPU memory (core is still stock).

Second Screenshot is at 1293/3602. Stable, but it runs too hot for me to consider keeping it there (70 on load on the GPU, plan is to add a second 680 in march/april so I want temps under 60). Memory and VRM's are well within range though so I'll leave the memory OC there I think !

SpannerMonkey

http://www.3dmark.com/fs/56521


FIRESTRIKE
3D Mark Score 9637

Graphics Score 12005
Physics Score 11377
Combined Score 3558

i7 3770k @4.3 Ghz
GTX 690 @915 / 1502
Air cooled with BeQuiet Dark Rock Pro 2



Seem to get worse scores if I put my overclockers back up to 4.6

W0LFMAN

Serephucus and SpannerMonkey

If you can upload screenshot like other, more than happy to up them on leader board.

ED E

You do know that'd be ridiculously easy to fake right?

W0LFMAN

ED E wrote: »

You do know that'd be ridiculously easy to fake right?

Sure, could easily do it myself.

But looking for a sense of honest Playmanship here, wouldn't make sense to Fake a score, would do nothing but ruin the integrity of the Boards PC overclocking forum members.

Could understand it there was a Prize in it. but as there is only Stats here, I'm sure they would be information useful in there own right.

what you think?

ED E

Exactly, so I dont see the point in the image checks, it'll just mean less people will bother.

Anyways, I'm wondering if I should go back and try and push my ram clocks

Serephucus

I only have the free version, and those bloody demos make the thing pretty damn long. I'll try run it later I guess. Personally I don't really see the need for all the info, name, etc. except maybe if other people are curious of specs.

SpannerMonkey

I have my boards user name in my link . And the link is direct to their records . Thats impossible to fake.
And the hassle of going uploading screenshots etc
I can change my name to boards.ie if you want

W0LFMAN

I Understand your comments fellas

Going to let a poll Decide.

Poll will be up for week.

What-ever get the majority vote, we be the rule.

Can't be fairer than that now boys can I really?

Serephucus

Might I suggest something like:

ICESTORM
Rank |Member Name|Score - Rig Specs
1.|Bloodbath|177,300 - (I5 3570k, 8g Ram, Radeon HD 7970)
2.|Deano12345|174,417 - (I7 2600k, 32g Ram, GTX 680)
3.|W0LFMAN|128,509 - (I7 990x, 24g Ram, GTX 480)
4.|ED_E|106,327 - (Quad Core Q6600, 8g Ram, Radeon HD 7870 )
5.|Marco_Polo|104,438 - ( AMD Phenom II X4 965, 8g Ram, Radeon HD 7850 )
6.||
7.||
8.||
9.||
10.||
11.||
12.||
13.||
14.||
15.||


For the main layout? Others can be adjusted if you copy my table into a post box then edit.

SpannerMonkey

W0LFMAN wrote: »

I Understand your comments fellas

Going to let a poll Decide.

Poll will be up for week.

What-ever get the majority vote, we be the rule.

Can't be fairer than that now boys can I really?

i dont think you should just be able to claim what you want , but if you provide a link to your score on their website with your boards user name on it surely thats enough ?

Deano12345

ED E wrote: »

Exactly, so I dont see the point in the image checks, it'll just mean less people will bother.

Anyways, I'm wondering if I should go back and try and push my ram clocks

Lemme know if you do and it makes a difference ! I reckon theres a little bit more left in this IMC, I'd go for 2133 if it helps my score

W0LFMAN

Serephucus wrote: »

Might I suggest something like:

ICESTORM
Rank |Member Name|Score - Rig Specs
1.|Bloodbath|177,300 - (I5 3570k, 8g Ram, Radeon HD 7970)
2.|Deano12345|174,417 - (I7 2600k, 32g Ram, GTX 680)
3.|W0LFMAN|128,509 - (I7 990x, 24g Ram, GTX 480)
4.|ED_E|106,327 - (Quad Core Q6600, 8g Ram, Radeon HD 7870 )
5.|Marco_Polo|104,438 - ( AMD Phenom II X4 965, 8g Ram, Radeon HD 7850 )
6.||
7.||
8.||
9.||
10.||
11.||
12.||
13.||
14.||
15.||


For the main layout? Others can be adjusted if you copy my table into a post box then edit.

Thanks Serephucus for the generous help. Compliments posted on main body.

SickBoy

Driver Version: 310.70
CPU: i7-2600 @ 4.1ghz
GPU: 1 x EVGA GTX670 FTW @stock

http://www.3dmark.com/3dm/96016


What's with the "Graphics Driver is not approved" warnings?

Deano12345

SickBoy wrote: »

Driver Version: 310.70
CPU: i7-2600 @ 4.1ghz
GPU: 1 x EVGA GTX670 FTW @stock

http://www.3dmark.com/3dm/96016


What's with the "Graphics Driver is not approved" warnings?

I'd imagine its to do with drivers that were released before/after this version of 3DMark was and the program hasn't been updated to register those drivers as legit, or something to that effect, since my 96's are Beta drivers and are picked up fine, but they would have been released just before 3DMark was

SpannerMonkey

will this do ?

I7 3770k
16GB Corsair Vengance
Gtx 690

Direct links (for a more zoomed in view)
http://i.imgur.com/yUmmxFH.jpg
http://i.imgur.com/ZFjxwRH.jpg
http://i.imgur.com/DQ2gM1J.jpg

[IMG][/img]

[IMG][/img]

[IMG][/img]

W0LFMAN

Updated. 06-02-13

(Very Nice Rig SpannerMonkey)
(Overclock CPU to 5Gz and throw in another GTX 690, I think you would be unbeatable)

SpannerMonkey

W0LFMAN wrote: »

Updated. 06-02-13

(Very Nice Rig SpannerMonkey)
(Overclock CPU to 5Gz and throw in another GTX 690, I think you would be unbeatable)

Haha cheers
. But nah i normally keep cpu at 4.6 but any more than that and it becomes unstable . And oddly i get better results on this at 4.5 , and as for a secong gtx 690 , not for quite a while they are a tad pricey ;-P
May go 2011 socket in time but thats about it :-)
its already hugely overkill for what i need

Serephucus

Here we go:

IS: 175793
CG: 19427
FS: 6300

3570K @ 4.6
GTX 670 @ 1042(1228 boost)/3402
16GB 1600MHz DDR3

I also think it might be a good idea to include CPU clocks. As most people are running similar things, it would give others a rough idea of how much a given overclock will help.

ED E

6th in every test, least mines consistent :P

dwighet

Were do I find this 3D mark??? Is it free?

SpannerMonkey

dwighet wrote: »

Were do I find this 3D mark??? Is it free?

There is a link to it in the first post ;-)

Serephucus

http://www.pcgamer.com/2013/02/06/3dmark-wars-nvidia-and-amd-go-head-to-head-on-our-test-rig-who-wins/

Bit of an interesting read. Seems like the weird behaviour with AMD cards hasn't been rectified quite yet, though this is a new benchmark. Doesn't say what drivers are being used, but seeing as it was published yesterday, I'd imagine they're up to date.

W0LFMAN

Humm, Can't seem to edit the top post any more..

Any Idea's?

BloodBath

Amd cards always seem to spike more in games if you look at any benchmarks that provide graphs.

I really don't mind though when i'm getting better than 680gtx performance for nearly half the price.

Damn you all with your HT processors.

I tried squeezing a bit more out of my clocks but couldn't get enough to improve the scores by much. This sapphire card doesn't run as cool as I was hoping but it's rock solid at 1200/6200 so I can't complain really.

I could probably get to 1300/6500 if I set the fan speed to manual 100% just for the benchmark.

Serephucus

It's a pity, but my 670 doesn't seem to be a great overclocker. I'm currently sitting at +125/+400, and I'm hitting the power limit. Thing is that the card never goes above 45°C, and thermal throttling doesn't kick in until 70. I have so much more headroom thermally it's killing me. I considered a voltage unlocked BIOS, but really it would only be for e-peen.

CPU clocks seem to help quite a bit in this. Obviously not nearly as much as GPU clocks, but more-so than in Vantage and 11.

BloodBath

It could be worth it. Voltage locks was one of the main things that put me off the Nvidia cards. That and the price.

You could probably get reasonably higher clocks. My card will only do about 1050/6000 at stock voltage. It will do 1250-1300 if I could keep it cooler. It starts pushing 70c at 1200 with my current reasonably relaxed fan profile to keep it quiet.

If you have that thing water cooled then you would be crazy not to. Does it have a dual BIOS?

Yeah the cpu clocks/threads are helping because of the physics tests in every benchmark. They seem to affect the score more in some of the benchmarks. The second test is heavily affected by it. I have 25% lower scores than some people with HT chips but my frames drop right down in the physics tests on that into the low 20's towards the end.

W0LFMAN

Guys, I think There is a time limit on editing posts (24 hours or maybe a bit longer) but after that, maybe moderator will have to do the editing from now on?

ED E

W0LFMAN wrote: »

Guys, I think There is a time limit on editing posts (24 hours or maybe a bit longer) but after that, maybe moderator will have to do the editing from now on?

Yep, there is.

Serephucus

BloodBath wrote: »

It could be worth it. Voltage locks was one of the main things that put me off the Nvidia cards. That and the price.

You could probably get reasonably higher clocks. My card will only do about 1050/6000 at stock voltage. It will do 1250-1300 if I could keep it cooler. It starts pushing 70c at 1200 with my current reasonably relaxed fan profile to keep it quiet.

If you have that thing water cooled then you would be crazy not to. Does it have a dual BIOS?

Yeah the cpu clocks/threads are helping because of the physics tests in every benchmark. They seem to affect the score more in some of the benchmarks. The second test is heavily affected by it. I have 25% lower scores than some people with HT chips but my frames drop right down in the physics tests on that into the low 20's towards the end.

I watercool for the silence first though, any performance benefits are a bonus, and really, the 670 does everything I want at present. I specifically bought it as a stop-gap card because I got twice the length to Step-Up. I have until June or something to upgrade, so I'll probably just jump to a 780 or Titan or whatever it ends up being. And no, no dual BIOS sadly.

marco_polo

Serephucus wrote: »

http://www.pcgamer.com/2013/02/06/3dmark-wars-nvidia-and-amd-go-head-to-head-on-our-test-rig-who-wins/

Bit of an interesting read. Seems like the weird behaviour with AMD cards hasn't been rectified quite yet, though this is a new benchmark. Doesn't say what drivers are being used, but seeing as it was published yesterday, I'd imagine they're up to date.

The latency issue is still there, they shored up the problem for some of the identified games in 13.1 but there were still a few weeks away from completing a rewrite the memory subsystem as a more general fix.

I suppose the silver lining for AMD card owners is that since this shows that since the driver team still haven't fully got to grips with the newer GCN architecture, there may still be a bit more to squeeze out of current cards in future driver updates.

BloodBath

Yep it's an ongoing fix. 13.2's aren't even officially out yet and there will be more latency fixes in future revisions.

I don't understand why people claim Ati driver's are ****. There's constant updates and fixes to any issues. Maybe not as fast as Nvidia but far from slow.

I wonder how it will impact their performance though when they do get to fixing the latency fully. There has been some small drops in frame rates from the initial fixes.

I notice the difference in the 13.2 betas though. Everything is a lot smoother. Could be a placebo but I don't think so. Is there any benchmark for testing frame latency?

ED E

13.2 fixed some issues for me with my 7870.

Have to agree, alwasy found ATI are on the ball for patching drivers till they get them just right.

ED E

Did some testing just to see

Base tests to get an idea of what I should get
Test # |Settings|Scores|URL
1.|Stock Clocks|106811|
|Misc Programs off|12790|
|MSE ON|4517|
|AB ON||http://www.3dmark.com/3dm/126114
2.|Stock Clocks|106456|
|Misc Programs off|12649|
|MSE ON|4519|
|AB ON||http://www.3dmark.com/3dm/126544
3.|Stock Clocks|105756|
|Misc Programs off|12784|
|MSE ON|4521|
|AB OFF||http://www.3dmark.com/3dm/127009
4.|Stock Clocks|106625|
|Misc Programs off|12812|
|MSE OFF|4518|
|AB OFF||http://www.3dmark.com/3dm/127618


Then two more with some tweaks to see what affect they have:

Test # |Settings|Scores|URL
5.|1150C/1250M|105745|
|Misc Programs off|12838|
|MSE OFF|4807|
|AB OFF||http://www.3dmark.com/3dm/128409
6.|Stock|106332|
|Misc Programs off|12781|
|MSE OFF|4518|
|AB OFF||http://www.3dmark.com/3dm/128820
|CataSettingsSetToNoAAEtc


Notes
Initial tests are stable/repeatable. Good for a benchmark and something Id expect from futuremark. Changes didnt increase things at all, except bumping clocks which only increased the Firestrike test, which makes sense. FPS is super high in the first tests so its probably not GPU limited.

First time I ran things it detected my CPU clock and memory, none of these tests did

Serephucus

AB?

ED E

Afterburner, should have put that in

Deano12345

Serephucus wrote: »

AB?

Afterburner I'd say !

Serephucus

Deano12345 wrote: »

Afterburner I'd say !

Hey, I was proud of myself for figuring out Security Essentials!

(Though thinking about it, not much else AB could be)