Post your benchmarks!

It is quite easy to find 3DMark scores of most laptop graphics cards, but actual gaming performance is often harder to find and, as benchmark resources such as Notebookcheck do not seem very popular here, actual user-contributed benchmarks can be much better. I'm promoting similar initiatives in my site and in a Spanish forum, so I think in a few weeks we can have a good amount of benchmarks.

In order to get results to compare many different laptops, I have chosen 4 demos that include an internal benchmark (click to go to the demo download page):

Crysis (1.8 GB) - Devil May Cry 4 (1.3 GB, the benchmark is less than 500 MB, can be downloaded here) - World in Conflict (1.2 GB) - X3: Terran Conflict (400 MB)

(Final games usually perform better than the demos, so to be able to make the comparison with people not owning the games, it is important to use the demos, and not the final games)

Excepting X3: Terran Conflict, the rest of the games seem to be optimized for Nvidia cards, this has to be taken into account when comparing laptops with graphics cards from different brand.

The settings to be used are (AA and filters off if not stated otherwise):

Crysis GPU benchmark (both DX9 and DX10, if possible; you can find useful the this benchmarking tool)

* 800x600, Low settings
* 1024x768, Medium settings
* Native resolution, High settings

Devil May Cry 4 (both DX9 and DX10)

* 800x600, Texture resolution Low, Shadow Quality Off, Quality Low
* 1024x768, Texture resolution High, Shadow Quality High, Quality High
* Native resolution, Texture resolution Super High, Shadow Quality Super High, Quality Super High

World in Conflict

* 800x600, Low settings
* 1024x768, Medium settings
* Native resolution, High settings

X3: Terran Conflict

* 1024x768, Texture quality Low, Shader quality Low, AA Off
* 1024x768, Texture quality Medium, Shader quality Medium, AA x2
* Native resolution, Texture quality High, Shader quality High, AA x4

If you are not in the mood of running all the tests, you should at least run the 1024x768, Medium setting ones, so that all the cards can be compared in at least 1 test. All tests have to be ran without overclocking (you can send later the results with overclocking to compare them).

To calculate your results (only average FPS) you can fill this worksheet. You may include screenshots of the results obtained.

I'm summarizing the results here: http://optimitza.cat/benchmarks/laptops

Here we go!

 

What is the download size of each of the demos?

 

They are very big... I updated the original post to show the size. I learned recently that in the US it's usual to have limited the amount of information you can download... It's not that way in Spain and other countries.

 

Dell Studio 15 (1537)

* ATI Mobility Radeon HD 3450, 256 MB DDR2
* Intel Core 2 Duo T5800 (2,0 GHz), 3 GB DDR2-800
* Monitor 1280x800
* Windows Vista Home Premium SP1 32-bit


Crysis DX9

Can't run it

Crysis DX10

* 800x600, Low: 32.9 FPS
* 1024x768, Medium: 8.9 FPS
* Native, High: 3.8 FPS

Devil May Cry 4 DX9, average of the 4 scenes

* 800x600, Low: 53.6 FPS
* 1024x768, High: 18.7 FPS
* Native, Super High: 13.5 FPS


Devil May Cry 4 DX10, average of the 4 scenes

* 800x600, Low: 33.8 FPS
* 1024x768, High: 13.1 FPS
* Native, Super High: 10.2 FPS


World in Conflict

* 800x600, Low: 49 FPS
* 1024x768, Medium: 14 FPS
* Native, High: 4 FPS


X3: Terran Conflict, average of the 4 scenes:

* 1024x768, Low: 43.0 FPS
* 1024x768, Medium, AA x2: 29.2 FPS
* Native, High, AA x4: 17.0 FPS

 

Well out of those games I only have Crysis at the moment.

I ran on windowed mode, and for native res-high settings I had to use 1280x720 since my screen is 1366x768 but it is not supported by the tool. Ran three times but only used the last part to write this post otherwise it would take a lot of space.


BENCH RUN- 29/04/2009 02:59:34 a.m. - Vista 64

DX9- results

Run #1- DX9 800x600 AA=No AA, 32 bit test, Quality: Low ~~ Overall Average FPS: 72.025
Run #2- DX9 1024x768 AA=No AA, 32 bit test, Quality: Medium ~~ Overall Average FPS: 47.33
Run #3- DX9 1280x720 AA=No AA, 32 bit test, Quality: High ~~ Overall Average FPS: 30.37


DX10 results

Run #1- DX10 800x600 AA=No AA, 32 bit test, Quality: Low ~~ Overall Average FPS: 72.21
Run #2- DX10 1024x768 AA=No AA, 32 bit test, Quality: Medium ~~ Overall Average FPS: 45.115
Run #3- DX10 1280x720 AA=No AA, 32 bit test, Quality: High ~~ Overall Average FPS: 24.44


Not bad, but not good either haha. I still can't believe that I am having trouble to run a 2007 game but well, it is still playable somewhat at native res, but wish I could run it better. It is the only game I have this much trouble running fine.

 

Thanks ryzeki! In some results your laptop got lower scores than a 9700M GTS:
http://optimitza.cat/benchmarks/laptops

The possible differences are:
-You ran them in windowed mode
-You used the game instead of the demo

I'm curious about another thing: can you select 1366x768 in-game? The benchmark tool has its limitations, I can't even select 1280x800, that is a very common resolution, so I ran the native tests without the benchmark tool.

 

haha i think you need to fix your original post. It says "Crysis (1.8 GB) - Devil May Cry 4 (1.3 GB, the benchmark is only 500 GB, maybe it can be dowloaded alone) - World in Conflict (1.2 GB) - X3: Terran Conflict ( 400 GB)"

i believe it should be 500MB and 400MB respectively. Unless im wrong and these are some crazy benchmarking tools. HAHA

 

3Dmark05 = 8400~
3Dmark06 = 4600~

No overclocking

 

/off
Dude, read the first post.
\off

BTW, can we post our FRAPS results on games, different from the demos above? I think it would pretty useful for guys, who own for example TF2 (like me ;P) and wonder on what settings what FPS they're going to make. I have a couple of TF2 bench's, made with FRAPS, while playing in a 10vs10 (+/- 2-3 dudes) match. Should I post them?

PS: TF2 = Team Fortress 2

 

Damn! Thanks, it's corrected now!

 

i managed 5820 in 3dmark 06 with my Studio xps 16 that has a HD 3670 in it.

 

The main body of benchmarks should be from the demos and settings above, so anybody could provide comparable results, but the more information, the better! So feel free to post your FRAPS benchmarks, 3DMark and others. Specially for the newer cards (ATI 4000 series and Nvidia "G.. nnnM")!

 

Well, I reran the tests to check if windowed mode affected, so I ran in full screen tests. Yes, on the game I can run at native 1366x768 and I do so on High settings because Medium looks much worse and does not offer much jump in performance--- I would say my average would be 28 fps on high and native, with lowest fps being 16 and highest about 40.

These are with 64 bits now.

Run #1- DX9 800x600 AA=No AA, 64 bit test, Quality: Low ~~ Overall Average FPS: 79.12
Run #2- DX9 1024x768 AA=No AA, 64 bit test, Quality: Medium ~~ Overall Average FPS: 54.255
Run #3- DX9 1280x720 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 34.37



Run #1- DX10 800x600 AA=No AA, 64 bit test, Quality: Low ~~ Overall Average FPS: 72.885
Run #2- DX10 1024x768 AA=No AA, 64 bit test, Quality: Medium ~~ Overall Average FPS: 45.015
Run #3- DX10 1280x720 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 26.86

my thoughts? Well, the scores presented by the Toshiba seem weird, becase being overall similar systems in specs, I have a stronger video card.

1) we have basically the same processor (I use 2.5Ghz speed) and Crysis is only bound on CPU at low res low settings. I should, at the very least, get the same fps here as the toshiba.

2) The 9700M GTS is the exact same card as my 9800M GS except it has less shaders. It has the same stock clocks and all. It should not be able to outscore my card regardless of the tests done.

3) Drivers used. It is true some drivers give extra performance in games, and so far the drivers I am using don't seem to help much in crysis, as I have ran the game better before.

4) OC, maybe me and the toshiba have the card OC, I flashed mine to GTS speeds so now it does not have any excuse to loose to a stock 9700M GTS, so maybe the toshiba user has a higher clock than mine?

5) The demo vs the final game. It is true that sometimes the demo gives a different performance than the final game, sadly I don't have the demo to confirm.

Anyways, I think drivers might be at fault here, since I am basically running a 9800M GTS card, and I might score worse than the 9800M GTS users from other notebooks like the Gateway FX. It is funny because it seems I have the exact same performance as a 9700M GTS, except in DX10 where I got a tad bit higher fps on medium.

 

I don't think we need 3DMark scores in here - as you made perfectly clear in your first post you can practically find such scores for every configuration out there. Raw game fps results on the other hand would be much helpful for users, imo.
Here's a quick bench of TF2 I made a couple of days ago:
*My config is in my signature*
Settings:
Resolution: 1280x800 (native for my laptop)
Level of detail: Everything is on "High"
Water detail: Reflect World
AA: None
FPS:
Min FPS: 44
Max FPS: 291 (the game pauses for a few secs after someone kills you - that's when the fps becomes above 100/200)
Avg FPS: 97.859

 

Well, I'm going to ask the user, just in case the weird results are his (hard to tell with so few data).

By the way, a Spanish user contributed benchmarks of the HD 4650, you may be interested: http://optimitza.cat/benchmarks/laptops

 

Specs below. Dox's 182.46. Stable OC @625/1625/1000 (played many games for hours with no crash).

Run #1- DX10 800x600 AA=No AA, 64 bit test, Quality: Low ~~ Overall Average FPS: 81.63
Run #2- DX10 1024x768 AA=No AA, 64 bit test, Quality: Medium ~~ Overall Average FPS: 57.88
Run #3- DX10 1280x720 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 39.29

The basic test is CPU bound, the second test may as well be with my card, and the third test is questionable. I recall actually getting higher FPS at 1280 x 1024 because of crappy scaling. 1680 x 1050 DX10 High is still in the 30's.

Here you go:

Run #4- DX10 1280x1024 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 39.38
Run #5- DX10 1680x1050 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 32.16

So tell me please how 1280 x 1024 runs faster than 1280 x 720?

And for future reference, my "playability" build got 36.4 frames with DX9 CCC level 4, 1680 x 1050, same clocks and driver. So yes, I can run crysis just fine >.>

 

If a GTX 280M couldn't play Crysis...

 

OK, so I added your results in the main table:
http://optimitza.cat/benchmarks/laptops
It'd be nice to have more benchmarks of both cards! I think I can get the benchmarks of an HD 3870 and a GTX 260M next week.

I can't find a way to insert a table in the forum, I would insert a copy of the table in the first post... Should I insert a screen capture as in the Spanish forum?

 

I'll run the tests on my FX 770m and HD 2600 over the weekend. If I dont have time to run the tests in both OCed and non-OCed, which would prefer? (I'd assume stock but wanted to double check).

EDIT: After looking at your table, might I suggest you change the "High" setting from native resolution to something more standarized because it wont help comapring cards when the resolutions are off by a large amount, say 1440x900 vs 1920x1200...If you want to capture the higher resolution settings without compromising the comparabilty of your chart you could add a 4th colum "custom max" or something which would be the highest native if it is higher than say 1440x900 which could be your "High" setting. That seems to be the new average for laptops with 14"+ screens (you're not going to find a decent card in anything smaller anyway".

 

FYI: the Devil May Cry 4 benchmark alone can be found here

 

That's a nice table, I'll run the rest of the benchmarks when I have the time, probably by the end of the weekend.

DMC4 has 4 different levels benched, which score are we using?

Here's Crysis DX9:

Run #1- DX9 800x600 AA=No AA, 64 bit test, Quality: Low ~~ Overall Average FPS: 103.85
Run #2- DX9 1024x768 AA=No AA, 64 bit test, Quality: Medium ~~ Overall Average FPS: 69.99
Run #3- DX9 1680x1050 AA=No AA, 64 bit test, Quality: High ~~ Overall Average FPS: 33.61

 

Heres crysis benchmark results for my m1730 (T9300, 8800m GTX SLI, 4GB RAM)

DX9

Run #1- DX9 800x600 AA=No AA, 64 bit test, Quality: Low ~~ Last Average FPS: 72.30

Run #2- DX9 1024x768 AA=No AA, 64 bit test, Quality: Medium ~~ Last Average FPS: 49.50

Run #3- DX9 1900x1200 AA=No AA, 64 bit test, Quality: High ~~ Last Average FPS: 32.71

DX10

Run #1- DX10 800x600 AA=No AA, 64 bit test, Quality: Low ~~ Last Average FPS: 65.98

Run #2- DX10 1024x768 AA=No AA, 64 bit test, Quality: Medium ~~ Last Average FPS: 46.11

Run #3- DX10 1900x1200 AA=No AA, 64 bit test, Quality: High ~~ Last Average FPS: 29.64

 

9800M GS 512MB GDDR3 @ 600-1500-800 (GTS clocks)

DMC4

For some reason, the DX9 part of the DMC4 demo is broken for me. I can't see the menu. Here's some DX10 runs, though:

800x600, Low settings
Scene 1: 223.94 fps
Scene 2: 186.00 fps
Scene 3: 221.04 fps
Scene 4: 164.87 fps
Rank: S

1024x768, High settings
Scene 1: 113.08 fps
Scene 2: 86.72 fps
Scene 3: 141.47 fps
Scene 4: 66.00 fps
Rank: S

1280x720, Super High settings
Scene 1: 99.00 fps
Scene 2: 68.23 fps
Scene 3: 129.02 fps
Scene 4: 59.43 fps
Rank: A

X3

Not sure if these went right, my fans were going up and down so I think Powermizer may have been screwing with my clocks. Anyway...

1024x768, Low Settings, 0xAA: 75.8 fps
1024x768, Medium Settings, 2xAA: 68.9 fps
1366x768, High Settings, 4xAA: 57.5 fps

X3 scales really well in terms of Visual Improvement:Requirement Increase

World in Conflict

Low & Medium runs in DX9, High in DX10 (WiC does this as part of its profile settings)

800x600, Low Settings: 85.0 fps
1024x768, Medium Settings: 61.0 fps
1366x768, High Settings: 33.0 fps

Crysis DX9

800x600, Low Settings: 82.76 fps
1024x768, Medium Settings: 56.91 fps
1280x720, High Settings: 33.165 fps

Crysis DX10

800x600, Low Settings: 74.085 fps
1024x768, Medium Settings: 44.12 fps
1280x720, High Settings: 25.775 fps

 

Stock please!

EDIT: After looking at your table, might I suggest you change the "High" setting from native resolution to something more standarized because it wont help comapring cards when the resolutions are off by a large amount, say 1440x900 vs 1920x1200...If you want to capture the higher resolution settings without compromising the comparabilty of your chart you could add a 4th colum "custom max" or something which would be the highest native if it is higher than say 1440x900 which could be your "High" setting. That seems to be the new average for laptops with 14"+ screens (you're not going to find a decent card in anything smaller anyway".[/QUOTE]

True. I think a 4th test is too much for most people to run and for the table to keep the legibility... Also the current mix of 16:9 and 16:10 resolutions makes difficult to select a higher standard resolution... I was thinking a lot about that, I finally went that route because similar laptops usually have similar screen resolutions.

 

Average of the 4 scores. I made a very simple worksheet to make the calculations.

 

Your worksheet link is broken.

 

Ooops! Thanks for pointing it out. It works now.

 

9600GT GDDR2 + P8600 2.4GHz

DMC4 Benchmark, DX9

* 800x600, Texture resolution Low, Shadow Quality Off, Quality Low

1 139.45
2 103.14
3 132.30
4 118.80

* 1024x768, Texture resolution High, Shadow Quality High, Quality High

1 56.37
2 41.76
3 63.82
4 36.57

* 1440x900, Texture resolution Super High, Shadow Quality Super High, Quality Super High

1 37.13
2 24.91
3 44.32
4 28.32

 

WТF??? I exited DMC4 benchmark to run DX10 version and now i can't find where it did install itself!

 

Well...I'm pretty sure all my results for the "high" setting will be almost useless since they're way off because of the resolution hit. Anyway, here are the results:

No Overclocks applied

m4400 (All tests performed in Windows 7 RC1 - 32bit, specs in my sig)

X3
Low: 67.3
Med: 61.7
High: 30.0


Crysis - DX9
Low: 98.8
Med: 39.4
High: 8.3

Crysis - DX10
Low: 77.8
Med: 29.4
High: 7.7

DMC4 - DX9
Low: 151.9
Med: 71.5
High:24.8

DMC4 - DX10
Low: 116.5
Med: 54.9
High: 25.4

WIC
Low: 79
Med: 45
High: 19

8510p (All tests performed in Vista Business SP1 - 32bit, 2Ghz T7300, 1.5GB RAM, HD2600 256MB, Native Res 1680x1050)

X3
Low: 63.1
Med: 47.0
High:21.6


Crysis - DX9
Low: 72.3
Med: 26.2
High: 5.4

Crysis - DX10
Low: 64.2
Med: 23.1
High: 5.0

DMC4 - DX9
Low:134.5
Med: 44.2
Hign:22.0

DMC4 - DX10
Low: 79.6
Med: 25.9
High:15.2

WIC - DX9
Low: 61
Med: 28
High:10

 

Great! Thanks!

 

Thanks! I need to know the amount of VRAM and what's your OS.
Your system RAM is @667 MHz?
Did you use the Crysis demo or the final game?

 

Dedicated is 512mb
Total is 2303mb

OS is Vista Home Premium 64bit

Yes, my ram is 667Mhz

I ran the benchmark in the full game. Patched to 1.2

Regards Jim

 

These scores ARE noticeably higher than those of the 9700M GTS!

 

The laptops with the most similar cards have run the tests at 1440x900, you could use this resolution also and I'll put a special entry in the table.

By the way, do you use Quadro or GeForce drivers?

 

New results.

 

Quadro Drivers. Will run another batch of tests later today.

 

Crysis benchmarks.

BTW, I'm filling out the chart, and the X3 demo only gave me one average number, not level-by-level results. How do I get those?

 

Demo or full game?

This is what you should get:

So you have the data for the 4 scenes, but you don't need to calculate the average yourself.

 

I used the Crysis demo. I'll try the X3 run again, but I don't remember seeing anything like that.

EDIT: Found it, forgot to check the saved files.

 

View attachment Benchmarks.zip

Filled out everything but DMC4 DX9, which won't work. Tweaked the Crysis part a little bit to fit in DX9/DX10 and also the fact that it only runs 3 loops and averages the last 2 by default.

 

DMC4 Benchmark, DX10

800x600, Texture resolution Low, Shadow Quality Off, Quality Low:

303.485

1024x768, Texture resolution High, Shadow Quality High, Quality High:

189.35

1680x1050 Texture resolution Super High, Shadow Quality Super High, Quality Super High:

107.58

----------

DMC4 Benchmark, DX9

800x600, Texture resolution Low, Shadow Quality Off, Quality Low:

290.7675

1024x768, Texture resolution High, Shadow Quality High, Quality High:

198.5725

1680x1050 Texture resolution Super High, Shadow Quality Super High, Quality Super High:

114.37

 

Here are my 1440x900 results for the m4400

DMC4 - DX9: 39.4
DMC4 - DX10: 38.1
X3: 35.3
Crysis - DX9: 13.8(1440x960)
Crysis - DX10: 11.7(1440x960)
WIC: 23

 

Thanks, chart updated!

 

Heres some more results for my M1730.
My results for DMC4 & X3 are both demo versions.

Devil May Cry 4 DX9

* 800x600, Low: 240.93 FPS
* 1024x768, High: 182.085 FPS
* 1900x1200, Super High: 121.79


Devil May Cry 4 DX10

* 800x600, Low: 258.88 FPS
* 1024x768, High: 199.56 FPS
* 1900x1200, Super High: 119.32 FPS


X3: Terran Conflict

* 1024x768, Low: 72.549 FPS
* 1024x768, Medium, AA x2: 72.321 FPS
* 1900x1200, High, AA x4: 58.125 FPS

Regards Jim

 

Chart's coming along nicely. I wonder why my benchmarks in Crysis aren't that great compared to the 9700M GTS, while everything else sits in the roughly 25-30% increase area.

If I have time in the next day or two, I'll flash back to GS clocks and get results for a stock G50Vt-X1.

Oh and I didn't run Crysis as a full game, so I don't know where those came from >_>

 

His are only better at lower resolution mate, its the same for me.
I'm very happy that I'm giving the Nvidia GeForce GTX 280M OC a run for its money. In DMC4 im getting a few higher fps at 1900x1200 than his at 1650x1080.
Thats probably because of drivers, though. Optimized ones for 280M card looks like it gonna be the one to beat.

Regards Jim

 

Yea, I'm using the new Nvidia beta drivers right now. Might reinstall DOX ones and see what happens.

I also think I suffer at lower resolutions from Nvidia scaling to a 16:9 screen.

 

I'm using those. Very happy with them at the moment. Saying that I havent benchmarked these games with any other driver.

I dont mind struggling at low res's I never use them to play anyway.

Did the benchmarks for WIC the full game patched to 1.009

* 800x600, Low settings 63 fps
* 1024x768, Medium settings 55 fps
* 1900x1200, High settings 41 fps

 

Don't have the time to so all that sheeet.

3dMark06 scores in sig though.