Anand has Farcry 1.2 patch and tests SM3.0

[jvd]

I believe they're not doing anything fishy... they were just sloppy with the said review (which imho is inexcusable for a "professional" site, but anyways)...

What are your feelings on the drop seen on the radeons from patch to patch.

 

[rainz]

Anand seems to be very frequent on making errors.

Yeh but i think it's not really an " errors ".. If you know what i mean.

It is a bit suspect that he didn't change his conclusion even though the new results invalidate it. :?

I know Dig i know :?

RainZ

 

[Kombatant]

I believe they're not doing anything fishy... they were just sloppy with the said review (which imho is inexcusable for a "professional" site, but anyways)...

What are your feelings on the drop seen on the radeons from patch to patch.

To be quite honest, you know what troubles me with that review? I have been closely looking at its ATI numbers for about five minutes. You know what I see? For every ATI card, the fps they get without the patch is ALWAYS the same as with the PS3 patch. For EVERY ATI card, for EVERY benchmark. Always the same. Even to the decimal! Everyone who has benched a card even once in his life knows that this is simply...well... impossible.

Any ideas on that other than the one I have in my head and it's not very flattering of Anandtech atm?

 

[digitalwanderer]

To be quite honest, you know what troubles me with that review? I have been closely looking at its ATI numbers for about five minutes. You know what I see? For every ATI card, the fps they get without the patch is ALWAYS the same as with the PS3 patch. For EVERY ATI card, for EVERY benchmark. Always the same. Even to the decimal! Everyone who has benched a card even once in his life knows that this is simply...well... impossible.

Damn good point, I missed that.

Any ideas on that other than the one I have in my head and it's not very flattering of Anandtech atm?

If'n you're thinking that they didn't really run the ATi cards on the patched version (or else never ran them on the unpatched version and just used the patched numbers as the unpatched ones) then I can't think of anything other than it was a one-in-a-bazillion longshot/coincidence. :?

 

[jvd]

kombant i wasn't looking at anand i was looking at the hexus.net ones where they show the performance drop of the ati cards between patches

 

[Demigod]

Sorry to be rude but can anyone answere my question (see my first post above) Why are objects missing in the desk missing on the ati shot, an ammo box missing on the nvsm2 shot and everything there on the sm3 shot?

http://www.anandtech.com/video/show...reviews/video/nvidia/farcrysm3/volcanoati.jpg
http://www.anandtech.com/video/show.../reviews/video/nvidia/farcrysm3/volcanonv.jpg
http://www.anandtech.com/video/show...reviews/video/nvidia/farcrysm3/volcanosm3.jpg

I know an a64 shouldnt but is the game dynamicly reducing the number of objects? If so doesnt that make the score even less usefull?

 

[jvd]

i don't think it is . It could just be a capture bug . I know once in a blue moon when i move a certian wya some times things pop in and out

 

[Demigod]

Thank you

 

[Kombatant]

kombant i wasn't looking at anand i was looking at the hexus.net ones where they show the performance drop of the ati cards between patches

Could be as simple as a driver quirk, same as the drop the 6800 exhibits at 10x7.

 

[Deathlike2]

To be quite honest, you know what troubles me with that review? I have been closely looking at its ATI numbers for about five minutes. You know what I see? For every ATI card, the fps they get without the patch is ALWAYS the same as with the PS3 patch. For EVERY ATI card, for EVERY benchmark. Always the same. Even to the decimal! Everyone who has benched a card even once in his life knows that this is simply...well... impossible.

Someone's being lazy.. that's for sure. It does seem more along the lines of improper benchmarking. At least, he should've noted it.

The problem with the whole benchmarking situation with NVidia and Far Cry on a whole is that the data is not entirely testable by the masses.

1) No official NVidia drivers on NVidia's website. Yes, you can obtain them when you buy the card (or from other sites). However, there is no consistency whenever a new driver is released for new hardware or a new game appears. Anandtech has this issue... from the GF6800 preview/reviews... they are using some form of the 60.xx drivers that are unlikely to be used by those people that have the cards on hand.

It seems to me that these "reviews" need to be done over (just benchmark-wise) when official release drivers come out. Otherwise we'll all be referring to "driver version X, that's only available to the reviewers" to prove that some hardware performs better than another. The problem is REPRODUCABILITY. If you can't reproduce the results, what good is your data?

2) Farcry 1.2 is not released. Not a big deal yet.

3) DX9c (official, non-beta) is not out. This is critical to the Geforce 6800's PS3 features(caps), is it not? It is not unlikely that performance will hinge on the final version of the API. The betas can give you an idea, but it isn't final. MS hasn't really given word to when it'll be out. You could bench with the betas... and it's not too bad an issue though.

The biggest problem is that NVidia has no drivers for the Geforce 6800 "officially". Last known drivers have been available 2 months ago... The closest to anything NVidia has released is "hidden" via the Quadro drivers where the 60.85 resides. It bothers me greatly as these issues to continue. Heck, even Futuremark certified the 60.85 drivers (and it went kinda off the radar). This practice Anandtech does is terrible. Really horrible.

 

[Pete]

Well now after reading this stuff about anandtech with the new results I wont go to that site anymore.

Well, you can visit (surely AT will get some exclusives due to its ginormous readerbase that smaller sites won't), but just do so with a skeptical eye (as with any site).

Kombatant, Derek probably just used the same numbers for both. Keep in mind that he's not changing patches between runs, just paths. He can't run the ATi cards in SM3.0 mode, AFAIK, so this would seem to be a non-issue, as he didn't show numbers between patches (though that would've been informative). I agree that he could've tidied up the review by including SM2.0 and 3.0 numbers in the same graph, just using different colors. Maybe their Flash graphs can't accomodate that, though.

 

[Kombatant]

Well now after reading this stuff about anandtech with the new results I wont go to that site anymore.

Well, you can visit, but just do so with a skeptical eye (as with any site).

Kombatant, Derek probably just used the same numbers for both. Keep in mind that he's not changing patches between runs, just paths. He can't run the ATi cards in SM3.0 mode, AFAIK, so this would seem to be a non-issue, as he didn't show numbers between patches (though that would've been informative). I agree that he could've tidied up the review by including SM2.0 and 3.0 numbers in the same graph, just using different colors. Maybe their Flash graphs can't accomodate that, though.

Yes, you are right. A more valid comparison though would be to do what Hexus did, which is compare the two patches.

 

[Jack_Tripper]

Now you ARE doing that with you're 6800 review...right????



Jack

O'm tupongg bkindfodled

Jack

 

[Brandon]

In case you missed that:

UPDATE: It has recently come to our attention that our 4xAA/8xAF benchmark numbers for NVIDIA 6800 series cards were incorrect when this article was first published. The control panel was used to set the antialiasing level, which doesn't work with FarCry unless set specifically in the FarCry profile (which was not done here). We appologize for the error, and have updated our graphs and analysis accordingly.

Glad to see they got their article updated with the correct numbers. I knew something was fishy as soon as I saw that they were using a setup similar to ours. From there determining what they did wrong wasn't hard.

Regarding X800 performance with the new patch and whether it declines or not, I dunno. When I first got the patch I ran some quick numbers with our moneybay demo and the 9800 PRO and noticed a negligible performance difference so went on with testing from there. I wouldn't recommend the 1.2 patch to X800 users anyway until the current IQ issue is fixed (unless you want to run at a lower resolution, it also isn't affected on flat surfaces).

As far as AT's comments on 3DMark, keep in mind that 3DMark 03 is purely a graphics benchmark and AT is more of a CPU/motherboard site so it's not surprising that they wouldn't like it.

 

[fallguy]

They still havent fixed the text part of their article. In the final words, it still says ATi has the lead in their two custom demos, until AA/AF are enabled, then NV takes over. Which is directly opposite of what their revised charts say. Why even bother if you arent going to fix it all.

 

[Pete]

It *is* a holiday weekend in the US, fallguy. Credit to him for updating the numbers on a Saturday, but you can give him the weekend to enjoy some outdoor grillin' before Anand sends him back into the benchmarking dungeon, no?

 

[Kombatant]

It *is* a holiday weekend in the US, fallguy. Credit to him for updating the numbers on a Saturday, but you can give him the weekend to enjoy some outdoor grillin' before Anand sends him back into the benchmarking dungeon, no?

I could get nasty now, by saying that he had time to update the graphs, but not the texts, but I will not...












hmm, i think i did nasty anyway



Btw nice forum title

 

[ANova]

I knew those AT numbers were bs from the very moment I laid my eyes on the graphs. They're review is FUBARed as far as I'm concerned.

 

[Stryyder]

I believe they're not doing anything fishy... they were just sloppy with the said review (which imho is inexcusable for a "professional" site, but anyways)...

What are your feelings on the drop seen on the radeons from patch to patch.

To be quite honest, you know what troubles me with that review? I have been closely looking at its ATI numbers for about five minutes. You know what I see? For every ATI card, the fps they get without the patch is ALWAYS the same as with the PS3 patch. For EVERY ATI card, for EVERY benchmark. Always the same. Even to the decimal! Everyone who has benched a card even once in his life knows that this is simply...well... impossible.

Any ideas on that other than the one I have in my head and it's not very flattering of Anandtech atm?

I noticed the same thing bothered me too but although it is highly impropable it is possible, or if there was a mistake perhaps he transcgribed his testing notes incorrectly into the chart.

Based on his comments the whole article has a 'rushed' feel to it, thats when people make mistakes.