nVidia's 53.03 and 52.16 drivers in 3dMark2003

[illuminati]

Alright.... time for a new thread for this topic . You all may have noticed that Dean at Overclocker Cafe did not notice a difference in 3dMark2003 scores.

Out of curiosity, I reran 3D Mark 2003 using the “approvedâ€￾ driver and ended up with the same results

Well, neither did I when using a 5600 Ultra on an Athlon FX-51 test system. With the 52.16 "Approved" driver, I received a score of 3273 3dMarks. With the 53.03 driver, I received a score of 3274. To me, that is not a difference. I also have the numbers for Fill Rate, Vertex Shader, Pixel Shader 2.0 and Ragtroll. So if someone requests that I post those numbers, or the FPS numbers for the individual game tests, I will be happy to do so.

What I'm hoping to get out of posting this is some explaination as to why the 52.16 driver is Approved, and the 53.03 driver is not. Worm, if you can give some sort of explaination, that would be great!... although, I understand if you are not able to.

Have any other members done any research between these two drivers to understand the difference between them? I'm very much interested because right now I'm thinking that any potential optimizations for 3dMark2003 are present in both the 52.16 and 53.03 drivers.

Thanks for everyone's help!

 

[xGL]

are both WHQL ?

 

[illuminati]

yes. Both the 52.16 and 53.03 drivers are WHQL certified.

 

[digitalwanderer]

Alright.... time for a new thread for this topic . You all may have noticed that Dean at Overclocker Cafe did not notice a difference in 3dMark2003 scores.

Out of curiosity, I reran 3D Mark 2003 using the “approvedâ€￾ driver and ended up with the same results

Well, neither did I when using a 5600 Ultra on an Athlon FX-51 test system. With the 52.16 "Approved" driver, I received a score of 3273 3dMarks. With the 53.03 driver, I received a score of 3274. To me, that is not a difference. I also have the numbers for Fill Rate, Vertex Shader, Pixel Shader 2.0 and Ragtroll. So if someone requests that I post those numbers, or the FPS numbers for the individual game tests, I will be happy to do so.

What I'm hoping to get out of posting this is some explaination as to why the 52.16 driver is Approved, and the 53.03 driver is not. Worm, if you can give some sort of explaination, that would be great!... although, I understand if you are not able to.

Have any other members done any research between these two drivers to understand the difference between them? I'm very much interested because right now I'm thinking that any potential optimizations for 3dMark2003 are present in both the 52.16 and 53.03 drivers.

Thanks for everyone's help!

Have you updated 3dm2k3 with the latest patch? Which version did you test it with?

 

[illuminati]

Sorry... Yes, we used build 340 of 3dMark2003.

 

[Hanners]

Have any other members done any research between these two drivers to understand the difference between them? I'm very much interested because right now I'm thinking that any potential optimizations for 3dMark2003 are present in both the 52.16 and 53.03 drivers.

This article I wrote may help:

nVidia and 3DMark03 - The Empire Strikes Back

These tests were run on an Asus GeForceFX 5900, and the performance differences are clear to see. I actually have a 5600XT in my possession now too, I might look into scores on that card at some point too.

 

[digitalwanderer]

I actually have a 5600XT in my possession now too, I might look into scores on that card at some point too.

Really?!? From where? Don't tell me someone gave a Bastard an nVidia card?!?!?

 

[illuminati]

Have any other members done any research between these two drivers to understand the difference between them? I'm very much interested because right now I'm thinking that any potential optimizations for 3dMark2003 are present in both the 52.16 and 53.03 drivers.

This article I wrote may help:

nVidia and 3DMark03 - The Empire Strikes Back

These tests were run on an Asus GeForceFX 5900, and the performance differences are clear to see. I actually have a 5600XT in my possession now too, I might look into scores on that card at some point too.

The next question I have: Is it possible that there is a difference between the WHQL certified 53.03 from nVidia and the 53.03 driver released from eVGA (which I am assuming is not WHQL certified)?

Do you still have the exact 53.03 driver you used in this comparison?

 

[Hanners]

The next question I have: Is it possible that there is a difference between the WHQL certified 53.03 from nVidia and the 53.03 driver released from eVGA (which I am assuming is not WHQL certified)?

Do you still have the exact 53.03 driver you used in this comparison?

As far as I'm aware the file versions between the 53.03 eVGA release and the final nVidia release were identical, and the fact that FutureMark didn't certify the driver seems to back that up.

I don't have the version of the driver I used in that article now I'm afraid, I've moved on to using the official 53.03 drivers now (apart from when benchmarking 3DMark03 of course! ).

 

[illuminati]

Hum... still smells fishy.

I'm sending you a PM to ask a few more questions. Thanks.

 

[Hanners]

I actually have a 5600XT in my possession now too, I might look into scores on that card at some point too.

Really?!? From where? Don't tell me someone gave a Bastard an nVidia card?!?!?

It wasn't free, that much is for sure.

 

[digitalwanderer]

I actually have a 5600XT in my possession now too, I might look into scores on that card at some point too.

Really?!? From where? Don't tell me someone gave a Bastard an nVidia card?!?!?

It wasn't free, that much is for sure.

Damn it, there goes a wild hope of mine....

 

[Hanners]

Damn it, there goes a wild hope of mine....

Look on the bright side - ByteMe will be pleased!

 

[digitalwanderer]

Damn it, there goes a wild hope of mine....

Look on the bright side - ByteMe will be pleased!

No he won't, he's NEVER pleased.

 

[Khedlar]

Hmm interesting. This for a ASUS 5900U

Results from early December, system is overclock in these tests.

52.16 - 4946
53.03 - 5719

Result from 10 minutes ago, system is not overclocked:

52.16 - 4371
53.03 - 4367

I don't know how to explain this. Both drivers where downloaded from last year as well as build 340.

 

[cthellis42]

Uh... Gremlins?

Heh... They the same 52.16's you used before, or did you download them from nVidia again? (Important thing to rule out.)

 

[Khedlar]

Yes they are the drivers I have stored in my driver file. No new downloads from anyone.

 

[Tim Murray]

well this is odd.

I should run benchmarks. 5950U, will do it once I get the chance.

 

[kyleb]

hum Khedlar, try setting your system clock back?

it is a longshot but who knows how nvidia is really playing things at this point.

 

[illuminati]

Very interesting indeed. I hope to have new numbers on my nVidia Reference 5900 Ultra later tonight as well.