PDA

View Full Version : Upcoming BFG FX 5700 Review


Pages : [1] 2

Clay
10-16-03, 09:40 PM
Hi everyone. My presence has been very limited lately due to work. However, I've got some fire in me for a review and I'll be needing it as I just have a few days to put it all together. :headbang:

I received a BFG Asylum FX 5700 Ultra 128MB today (this is the NV36). I cannot divulge any specifics about it until after 9AM on October 23, so please don't ask.

Anyway, I'd like to hear from everyone as to what you'd like to see in the review. I have a BFG Asylum FX 5900 Ultra 256MB and an ATI Radeon 9800Pro AIW 128MB for comparison. Due to a tight deadline (I need this to be ready to post on the 23rd) I would like to cut out any "fluff" benchmarks. I want to focus on a select few (2 to 3 tops) and then include some details IQ comparisons. So, having said that, fire away with any suggestions. I won't be able to include all of course but I'll do my best.

Also, I'll probably "freeze" including any suggestions by tomorrow evening so that I can crank out the review from that point on.

Thanks!

TheTaz
10-16-03, 09:45 PM
Originally posted by maxpower
Anyway, I'd like to hear from everyone as to what you'd like to see in the review.

An extremely in depth "driver cheat" investigation.

And, no.. I'm not a fanATic... just a regular consumer that wants to KNOW.

Thanks,

Taz

Clay
10-16-03, 09:50 PM
Originally posted by TheTaz
An extremely in depth "driver cheat" investigation.

And, no.. I'm not a fanATic... just a regular consumer that wants to KNOW.

Thanks,

Taz
Extremely in depth driver investigations take a lot of time. Unfortunately I just don't have time for that kind of effort (I'm limited to about 2-3 hours an evening). That's a great idea and one I'd really like to pursue in the near future.

TheTaz
10-16-03, 09:53 PM
Originally posted by maxpower
Extremely in depth driver investigations take a lot of time. Unfortunately I just don't have time for that kind of effort (I'm limited to about 2-3 hours an evening). That's a great idea and one I'd really like to pursue in the near future.

Well, give what attention you can, on it. Any info is appreciated. :)

Regards,

Taz

{Sniping}Waste
10-16-03, 09:56 PM
1 of the 3 test needs to be a DX9 test.(Not GunMetal, thats not a real DX9 test with only one VS 2.0 shader).

Hellbinder
10-16-03, 10:00 PM
Ok how about benchamrks and basic IQ observations from at least one game from each Genera(sp?)..

Like

1 FPS game i suppose UT2003 is ok.
1 Racing game
1 Flight sim game
1 RTS game like War Craft III
1 Sports game
1 Standard RPG
1 MMO RPG like perhaps SWG
1 an action type game

Rogozhin
10-16-03, 10:10 PM
For your flight sim benchies PLEASE run Forgotten Battles (IL-2 STurmovik FB).

Run with stencil buffer on and run "the black death" track (you'll need at least a 3.0+ ghz P4 and 1 gig of ram to make it GPU dependant with "perfect" settings-and do high res high AA AND AF.

It has pixel shaded water and is one of the most intensive benchies I've ran (aquamark and 3dmark are great for the cutting edge but FB brings systems to their knees even without aa and af).

thanks ;)

rogo

jimmyjames123
10-16-03, 10:16 PM
What drivers will you be using? Is NVIDIA going to have some new Det's out before your review comes out?

Clay
10-16-03, 10:21 PM
To {Sniping}Waste and Hellbinder...

I don't have Halo or TRAOD for DX9 benching. :( Other suggestions?

Hellbinder...ew..that'd be eight (8) games you listed there. Just can't swing that. I'll probably just be testing at 1024 and 1600 (skipping 1280 or an other middle ground) so that'll speed things up a bit but 8 is still a lot. I could swing IQs from more games though which is probably a good idea.

Thanks for all the feedback so far guys! Also, I'll probably "freeze" including any suggestions by tomorrow evening so that I can crank out the review from that point on.

Clay
10-16-03, 10:32 PM
To Rogozhin and jimmyjames123...

Unfortunately I do not have IL-2 STurmovik FB. I agree that it is a great benchmark. I have the IL-2 demo but that's it.

I can't say anything specific about the drivers. I received a printed note in the box though from BFG suggesting that the drivers on the retail CD not be used for testing. Instead, they suggested that I use a newer set of WHQL drivers. They also indicated that the greater majority of their products will soon contain these newer WHQL drivers.

ChrisW
10-16-03, 10:33 PM
Custom demos! This will let us know if any optimizations apply to the entire game or just to the benchmark. If is applies to the entire game, I really don't have a problem with it.

Set up a digital camera and focus it on a section of the screen. Stand still in a game and take a picture. Now use the screen shot feature of your game (or Hypersnap) and take a picture of the captured image. Now compare the image taken by you standing still on the screen to the picture of the screen capture image just to prove nVidia is not detecting when the end user is taking a screen shot and "enhancing" the image. This is the one allegation that must be disproved immediately! I really don't understand why absolutely nobody is willing to investigate this allegation.

ragejg
10-16-03, 10:33 PM
Humus' demos...

I think they may have some benching usage...

The Baron
10-16-03, 10:38 PM
use my UT2003 demo

http://www.nvnews.net/reviews/sapphire_radeon_9600_pro/nvnews.zip

think that URL works... and why hasn't this been stickied yet?

Clay
10-16-03, 11:02 PM
Originally posted by ragejg
Humus' demos...

I think they may have some benching usage...
Sorry ragejg...I dumb or sumthin'...what is Humus?

Clay
10-16-03, 11:03 PM
Originally posted by The Baron
use my UT2003 demo

http://www.nvnews.net/reviews/sapphire_radeon_9600_pro/nvnews.zip

think that URL works... and why hasn't this been stickied yet?
Got it, thanks! I also have my own UT2K3 demo from a while back that's pretty intensive I plan on using. I'll use yours too Tim.

Clay
10-16-03, 11:04 PM
Originally posted by ChrisW
Custom demos! This will let us know if any optimizations apply to the entire game or just to the benchmark. If is applies to the entire game, I really don't have a problem with it.

Set up a digital camera and focus it on a section of the screen. Stand still in a game and take a picture. Now use the screen shot feature of your game (or Hypersnap) and take a picture of the captured image. Now compare the image taken by you standing still on the screen to the picture of the screen capture image just to prove nVidia is not detecting when the end user is taking a screen shot and "enhancing" the image. This is the one allegation that must be disproved immediately! I really don't understand why absolutely nobody is willing to investigate this allegation.
I feel a bit "dim" tonight. ;) Are you serious about this or just being sarcastic? :) I just really can't tell for some reason.

The Baron
10-16-03, 11:04 PM
Originally posted by maxpower
Got it, thanks! I also have my own UT2K3 demo from a while back that's pretty intensive I plan on using. I'll use yours too Tim.
Mine is more intensive than any custom demo I've ever seen... seems like it needs at least a 1.8Ghz box to be not CPU-dependent, though. But, after that point, it's pure GPU-killing fun.

The Baron
10-16-03, 11:05 PM
Originally posted by maxpower
I feel a bit "dim" tonight. ;) Are you serious about this or just being sarcastic? :) I just really can't tell for some reason.
No. It's real. Valve mentioned a screenshot hack, and people are freaking out about it.

Clay
10-16-03, 11:09 PM
Originally posted by The Baron
No. It's real. Valve mentioned a screenshot hack, and people are freaking out about it.

Ok, I was aware of the claim by Valve but taking a picture of your monitor with digital camera doesn't seem to be like a "stable" comparison. I'll experiement though and definitely include this for at least one screenshot comparison.

Rytr
10-16-03, 11:18 PM
ragejg correct me if I am wrong but this is what he was referring to:http://esprit.campus.luth.se/~humus/


Humus posts here from time to time. Here's a thread on his work:http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=18730&highlight=humus

Clay
10-16-03, 11:21 PM
Originally posted by Rytr
ragejg correct me if I am wrong but this is what he was referring to:http://esprit.campus.luth.se/~humus/
Thanks stranger! :D Looks very interesting!

ChrisW
10-16-03, 11:35 PM
Originally posted by maxpower
Ok, I was aware of the claim by Valve but taking a picture of your monitor with digital camera doesn't seem to be like a "stable" comparison. I'll experiement though and definitely include this for at least one screenshot comparison.
I was thinking about taking a picture of a section of the screen just to see if there are some kind of visual anomalies, like fsaa being totally different or texture differences. I don't know what the results would be but it would certainly ease my mind over the subject if someone tried some type of image comparison that did not solely rely on screen captures. Call me crazy, but people are going to always have this question in the back of their minds unless someone does something to disprove it.

Clay
10-17-03, 12:02 AM
Originally posted by ChrisW
I was thinking about taking a picture of a section of the screen just to see if there are some kind of visual anomalies, like fsaa being totally different or texture differences. I don't know what the results would be but it would certainly ease my mind over the subject if someone tried some type of image comparison that did not solely rely on screen captures. Call me crazy, but people are going to always have this question in the back of their minds unless someone does something to disprove it.

Well here is my concern about doing this. Check out the attached image (it's a GIF so no compression, PNG was too large to post). It's a cropped area of a 1024x768 image that I took (in macro mode) of my desktop. The resolution isn't important here, but rather the fact that I was in macro mode you can see much more detail (actually degrades quality in that you get kind of a "screen door" effect). When I take pictures of the entire desktop the end result is really poor. I assume that this may be related to my camera somehow. I have a Nikon Coolpix 775 (2.1 MP).

At any rate, I don't think an apples-to-apples comparison can be drawn with this approach. I really like the idea though and am open to any suggestions. (I tweaked my exposure settings, white balance, etc, etc and never got what I felt was a photo that could realisticly be compared to an actual screenshot.)

gstanford
10-17-03, 01:44 AM
A suggestion for the screenshot/photograph comparison:

What you could do is photograph the screen as suggested, then do a screen dump, display the screen dump in irfanview or similar and photograph that also (not moving camera or screen in between times - this will require a tripod).

The results should them be comparable regardless of what the camera does to the picture quality.

Paul
10-17-03, 08:27 AM
Originally posted by gstanford
A suggestion for the screenshot/photograph comparison:

What you could do is photograph the screen as suggested, then do a screen dump, display the screen dump in irfanview or similar and photograph that also (not moving camera or screen in between times - this will require a tripod).

The results should them be comparable regardless of what the camera does to the picture quality.

So long as you keep the camera and monitor in the exact same position, this solution would be workable. I'm not sure you're going to be able to see much difference given the poor quality of pictures taken, but it's worth a try.