Last Updated: 5/31/09
Anything and everything related to the GTX 280M goes here!
Specifications for the GTX 280M:
*55nm fabrication
*754 Million Transistors
*128 Unified Shader
*585 Mhz Core Clock
*1463 Mhz Shader Clock
*800/950 Mhz Memory Clock
*75 Watt Power Consumption
*DX10, Shader Model 4.0 Support
*SLi Compatible
It's important to note their are 2 types of the GTX 280M, MXM 2.1 and MXM 3.0. Memory clock speed for 2.1 is 800 mhz stock, while memory clock speed is 950 mhz for 3.0. Clevo M570TU/ETU uses the MXM 2.1 version, Clevo D900f and M980nu use the MXM 3.0 version. The M980nu supports SLi GTX 280M.
Now that I have equipped my NP5797 with the fastest mobile CPU available, the 2.53 ghz QX9300, I am ready to give the GTX 280M a thorough test. Here are the conditions at which I ran all of the benchmarks:
*Nvidia's latest 185.85 drivers (with native 280M support).
*QX9300 is overclocked to 2.93 ghz.
*I am using Oile's overvolted 280M bios, for 1.05v.
*The 280M is clocked at 625/1625/1000 for every benchmark, unless noted otherwise.
*The resolution is 1680 x 1050 for every benchmark, unless noted otherwise.
*64 bit when applicable
*All retail games updated with latest patch
Why only overclocked results? Because I support extracting every bit of potential from your system while retaining stability. That's what this thread is all about! I may eventually add stock clocks if it's deemed necessary, but that takes time.
Crysis
GPU Benchmark
DX9 Medium: 58.065
DX9 High: 30.765
DX10 Medium: 50.085
DX10 High: 31.73
DX10 Very High: 18.68
![]()
CPU Benchmark
DX9 Medium: 59.55
DX9 High: 31.75
DX10 Medium: 46.64
DX10 High: 30.18
DX10 Very High: 19.155
![]()
Crysis Warhead
DX9 Mainstream: 61.41
DX9 Gamer: 32.12
DX9 Enthusiast: 20.94
DX10 Mainstream: 48.48
DX10 Gamer: 29.11
DX10 Enthusiast: 17.89
![]()
Far Cry 2 (Ranch Long)
DX10 Ultra 0x AA: 45.30
![]()
DX10 Very High 4x AA: 42.32
![]()
Devil May Cry 4 Demo
DX10 Super High Quality 0x AA
Scene 1: 116.49
Scene 2: 82.97
Scene 3: 143.37
Scene 4: 82.00
Average: 106.20
![]()
World in Conflict Demo
High Settings: min 21 max 66 avg 37
![]()
Very High Settings: min 16 max 50 avg 28
![]()
Grand Theft Auto 4 (with ultimate graphic's tweak)
Built-in benchmark
maximum settings: 39.16
tweaked settings: 44.21
FRAPS 10 minutes of gameplay, second island
maximum settings: min 14 max 43 avg 25.945
tweaked settings: min 19 max 104 avg 29.798
Fallout 3
FRAPS 10 minutes of gameplay, roaming in the wasteland
Ultra Settings 4x AA 15x AF
min 0? max 63 avg 42.78
Bioshock
FRAPS 10 minutes of gameplay in "Smuggler's Hideout"
DX10 High Settings
min 48 max 167 avg 91.417
Call of Duty: World at War
FRAPS 10 minutes of gameplay on "Eviction", many deaths lol, and a fallout 3 exe in the background by accident
Maximum settings, 4x AA 16x AF
min 21 max 92 avg 48.525
S.T.A.L.K.E.R Clear Sky
FRAPS 10 minutes of gameplay, pitch black in the Cordon, moving with a stalker patrol
DX10 Enhanced Mode, maxmium settings besides Suny Rays on low and 0x AA
min 13 max 73 avg 39.014
Left 4 Dead
FRAPS 10 minutes of gameplay on Dead Air, part 1
Very high settings, 4x AA 16x AF
min 58 max 125 avg 85.076
-
-
thank you for your efforts
-
Some corrections:
the M570TU is also confirmed to support the card. XoticPC has confirmed the option to purchase the card for the system.
You may want to have a mod change your thread title if you want it to be the catch all GTX 280m thread. Something like " *** Official GTX 280M Thread *** " Instead of making it all about OCing and benching.
If it is to be the catch all GTX 280m thread, you may want to remove speculations. I like to keep things in a wiki-like format.
Since NBR has incredible SEO, its possible this thread could become one of the top results for those who search google for GTX 280m.
Thanks for the thread. Sorry for bickering. -
I am curios on want VRAM the 280M hold. Maybe they've upgrade to 0.9 timings and we can break the 1 Ghz stable freq. clock.
EDIT: I apologise for the triple post. My internet had some problems and it accidentally triple posted. -
Hi
Is it possible to change my 9800 M GTX in to a GTX 280M?
An where can I by a GTX 280M?:
Ich have a: QGX7; Intel® Core 2 Quad Q9650 (4*3,00 Ghz); 9800M GTX, 4Gb DDR2; Vista Black Edition 2009, -
New status: Phase 4 (Order Shipped)
Some additional notes placed on your order:
Your order with Xotic PC has shipped!
You can track your packages below.
Thanks
Xotic PC
Estimated Wednesday -
why 1680x1050?
-
Because I don't need 1920 x 1200 for regular apps, and games will run optimally at native res for a longer time than if I chose 1920 x 1200.
-
So...anyone got the GTX 280m yet? -
Should have mine Tuesday. Supposedly it was one of the first two that shipped from PCtorque.
-
Wednesday... ordered a performance mouse and pad to go with it
-
I also will have mine wed, and I have the day off! Benchmarks to follow.
Gaming Rig:
Sager NP5797 | XP SP3 x86 | 17" 1920 x 1200 | 1GB GDDR3 GTX 280M | 3.06 GHz X9100 | 3GB 1066 DDR3 | 320GB 7200 RPM
Travel Rig:
Sony VGN-SZ650N/C | XP SP3 x86 | 13.3" 1200 x 800 | 8400M GS | 2.2 GHz | 160GB 7200
Folding team TFL # 160821 -
ohhhhh snap crackle POP
lol -
Congratulations!
Specs look very nice, looks like something I would get.
No Quad? -
-
wow, i wonder when i can ordered the GTX 280m and install it on my lappy!! cant wait to see the benchmark!! wuhuu!!
-
If any of resellers are to be believed, you'll have to wait in line for another couple weeks or so with the rest of us. -
I'll have some benches (3dmark, vantage, crysis, far cry 2) ready this afternoon. I'll add more like warhead and world at war late tonight, and maybe try overclocking. That is if it delivers on time and someone at the house signs it
I can do more than (but no higher than) 1680 x 1050 if requested. For crysis I'll do dx9 and dx10 very high, high, ccc level 4+5. -
-
Put this in the wrong thread, sorry.
Here is my first shot at 3d Mark 06 with the drivers it shipped with (6.14.11.7927)
http://service.futuremark.com/compare?3dm06=10657923
The only oddity is that the 8 Triangles test sucked at 11.6 MTriangles/s as compared to my 8400M GS that hit 26.0 MTriangles/s.
So I'll chalk that up to a driver issue and I'll be trying out a new driver today. Any suggestions?
Dave -
Wow, nich score, I want one too.....
-
Thanks for that comparison.
-
I also just realized that the drivers I have loaded are for Vista, not XP.
-
I loaded up the 6.14.11.7948 drivers and here is my lower score
http://service.futuremark.com/compare?3dm06=10661129
So, for now, stick with the OEM drivers.
Edit: Just figured out that since I was running RM clock the cpu was only at 2.93 for this test. -
This test was with 1x CPU overclock, which you have to change in the bios.
http://service.futuremark.com/compare?3dm06=10661921
Link to screen shot, also you can see the max temps and CPU overclock of 33%.
Here is the 2x CPU overclock.
http://service.futuremark.com/compare?3dm06=10662174
With the same screen shot as above, but different #s.
-
Wow,
I was relly hoping the 280M would beat 3870's in crossfire which is getting 15,000 for most people. I guess I have to give up my desire to get the 280M and instead get the 4870's when they come out. , I really wanted to get back to Nvidia.
Nice scores though! -
Interesting to see ORB report the graphics card as a GTX 180M rather than the 280M. Seems like a last minute decision by nVidia marketing to combat its competition. I don't think it would have sold well if they went with the GTX 180M.
-
Yeah I noticed that too. It did recognize the drivers as 280 though.
-
EDIT: Just noticed that ORB even recognized the 179.27 run as 180M. Hmmm -
Ok, finally typing on my new NP5797! Long day. Anyway, I'm hungry, I'll pop in the crysis disk and drop 10GB of stuff off my external while this updates and I grab some grub. I'll have benches up tonight. Any particular requests?
-
I guess the main ones are 1680x1050 high and very high as well as 1920x1200 high and very high.
-
Request: Find the max overclock of your GPU.
-
That's the goal by the morning
-
64 bit stock drivers and rivatuner are choking so I'm installing ntune
-
the internal crysis (NOT crysis warhead) benchmark utility (Benchmark_GPU batch file) that came with the game itself is the ONLY thing that matters. it is the ultimate benchmark and you must run it in the exact same resolution and settings as what i am posting below to compare apples to apples. make sure the settings update properly after you change them in the game's menu before exiting to run the Benchmark_GPU batch file. i post sli scores for referance. also forgot to mention, make sure you patch the game up before you run the benchies.
8800M GTX card(s) / crysis internal benchmark utility full screen mode Benchmark_GPU batch file;
800x600 all high dx9 = 49.16 fps avg. single gpu
1280x960 all high dx9 = 31.91 fps avg. single gpu
1440x900 all high dx9 = 29.62 fps avg. single gpu
1680x1050 all high dx9 = 23.37 fps avg. single gpu
1920x1200 all high dx9 = 18.86 fps avg. single gpu
800x600 all high dx9 = 43.50 fps avg. sli enabled
1280x960 all high dx9 = 41.67 fps avg. sli enabled
1440x900 all high dx9 = 38.56 fps avg. sli enabled
1680x1050 all high dx9 = 35.19 fps avg. sli enabled
1920x1200 all high dx9 = 30.28 fps avg. sli enabled -
ok, I'm going to verify this with yet another run, but dx10 high is slightly slightly faster than dx9 high. 28.07 vs 28.18. I honestly thought there'd be a 5~ fps difference in favor of dx9?
Also, max temp recorded so far 70c on the gpu, 40c on the p9600 running these benches. I am not even using the zalman...
28.225... DX10 > DX9 -
a variation of about 1 fps per completed bench run is normal, regardless of settings.
-
But shouldn't DX9 outperform DX10? Or de we finally have a card and drivers that took care of that issue?
EDIT: 15.885 on very high dx10 1680 x 1050, max GPU temp now 73c
EDIT 18.975 on CCC Lvl 5 dx10 1680 x 1050 -
DX does not matter in this game, only the video settings do. for simplistic reasons just stick with dx10 as that is the only way you can enable the very high settings with modding the congif files.
-
Tried 625/1550/950 and the driver crashed near the end of the first loop. Going to try "stock" 583/1450/950 now.
Well good news, 583/1450/950 ran solid. 18.975 went to 20.5 for the CCC Level 5 test, DX10 1680.
Going to try 3dmark06 now. -
how about some benches at "playable" FPS? like resolutions with 1440x900 and under?
-
@600/1500/950
DX10 1680 x 1050 CCC Level 4: 29.87 FPS
DX10 1680 x 1050 High: 30.73 FPS (2.5 above stock clocks) -
i have one same spec as yours coming in. went with the 1920 tho, kinda regret it now -
In the 11000's
-
is this the crysis benchmarks ???
A Falling Cat named Spangky
9 Lives only -
Yes, and CCC is an optimized config. Level 4 is an optimized very high mode but with less post processing, performs like high but looks better. Level 5 is an optimized very high mode with everything on, performs better than stock but still very demanding.
Just ran Far Cry 2, ultra settings on DX10, 1680 x 1050, no AA, 36.05 FPS. -
thats not very good
them GT725 from MSI is getting slightly better if not the same results ...
I am dissapointed ...
A Falling Cat named Spangky
9 Lives only -
-
A Falling Cat named Spangky
9 Lives only ... -
The GTX 280M Overclocking and Benchmarking Results Thread
Discussion in 'Sager and Clevo' started by anothergeek, Apr 10, 2009.