If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
FX5200 and fog problem
Hello,
I have a Asus card with nVidia FX5200, PC2700 memory and Athlon XP 1700+ with WinXP+DX9 and 44.67 driver. I am having problems in games with fog or smoke rendering. While the game runs smoothly at 1280x1024x32 and all other effects (lens flare, rain, water animations) are just fine, then whenever there is a bit smoke and I look through it, then I get around 3-5 fps. As soon as I look away from the smoke I have again normal FPS. I have not detected this behaviour on any other effect, just the fog. I have seen this issue on Battlefield 1942 patch 1.4 + road to rome and on GTA Vice City. Any ideas wether this is expected behaviour or a problem somewhere ? (I don't remember having this problem with my GF4MX440 and this was a DX7 card). Any help is appreciated, Mario |
#2
|
|||
|
|||
On Fri, 18 Jul 2003 14:52:19 +0300, Mario Kadastik
wrote: Hello, I have a Asus card with nVidia FX5200, PC2700 memory and Athlon XP 1700+ with WinXP+DX9 and 44.67 driver. I am having problems in games with fog or smoke rendering. While the game runs smoothly at 1280x1024x32 and all other effects (lens flare, rain, water animations) are just fine, then whenever there is a bit smoke and I look through it, then I get around 3-5 fps. As soon as I look away from the smoke I have again normal FPS. I have not detected this behaviour on any other effect, just the fog. I have seen this issue on Battlefield 1942 patch 1.4 + road to rome and on GTA Vice City. Any ideas wether this is expected behaviour or a problem somewhere ? Expected.........fog-effects hog GPU power. (I don't remember having this problem with my GF4MX440 and this was a DX7 card). Any help is appreciated, Mario The FX5200 is a slow card. To get an exact comparison you need to prune the effects back to those available on the MX440. Use NVtweak or similar. You probably should have considered a FX5200Ultra or better for your chosen screen-resolution and game software (and your FPS tolerance- threshold).. You also may have purchased the 64-bit data-path variant of the FX5200 (non-Ultra).............same memory capacity (128Mbytes or 64Mbytes) but less-expensive to manufacture.......... 64-bit data-path == memory bandwidth 3.2Gbytes/sec 128-bit data-path == memory bandwidth 6.4GBytes/sec Quite a few FX5200 (non-Ultra) manufacturers omit the memory bandwidth on their specs and I challenge you to see that figure on the retail box. Deliberate ?? Do not confuse 128-bit processing with 128-bit data-path. If you do not see a clearly-stated memory bandwith spec. on a manufacturer's data sheet, watch out.......... BTW, all FX5200ULTRA boards that I have come across clearly boast their memory bandwidth, normally around 8.0Gbytes/sec --- and thus definitely 128-bit data-bus, with the higher memory-clock speed. Well-informed as to desired performance and actual video card specs before purchase makes for greater happiness after purchase............ John Lewis |
#3
|
|||
|
|||
This is a known problem in Vice City, especially when frame limiter is off.
"Mario Kadastik" wrote in message ... Hello, I have a Asus card with nVidia FX5200, PC2700 memory and Athlon XP 1700+ with WinXP+DX9 and 44.67 driver. I am having problems in games with fog or smoke rendering. While the game runs smoothly at 1280x1024x32 and all other effects (lens flare, rain, water animations) are just fine, then whenever there is a bit smoke and I look through it, then I get around 3-5 fps. As soon as I look away from the smoke I have again normal FPS. I have not detected this behaviour on any other effect, just the fog. I have seen this issue on Battlefield 1942 patch 1.4 + road to rome and on GTA Vice City. Any ideas wether this is expected behaviour or a problem somewhere ? (I don't remember having this problem with my GF4MX440 and this was a DX7 card). Any help is appreciated, Mario |
#4
|
|||
|
|||
Hello,
John Lewis wrote: Expected.........fog-effects hog GPU power. Well shouldn't lens flare and a lot of polygons also hog gpu power? I'm not having any fps slowdown while fightinh with tens of japanese that have a lot of tanks and naval support with explosions and lens flare effects (sunset and so on). But when I look quite near at a tank that is smoking, then I get 3-5 fps. That shouldn't be GPU performance problem. (I don't remember having this problem with my GF4MX440 and this was a DX7 card). The FX5200 is a slow card. To get an exact comparison you need to prune the effects back to those available on the MX440. Use NVtweak or similar. where might I get NVtweak? You probably should have considered a FX5200Ultra or better for your chosen screen-resolution and game software (and your FPS tolerance- threshold).. Well actually the situation was that I had my MX440 and was quite happy with it (normal FPS in games and quite enough features to play games currently available. Also did play GTA Vice City and BF1942 and don't remember having problems with smoke). But when it got fried around a month ago (I got periodic 10sek pauses while gaming and so every 30-60 sek) I decided not to replace it with the same card but to do an upgrade that wouldn't cost anything big. My only necessity was that the card should have tv-out (quite standard these days) and would be better than the one I had. As I practically didn't have to pay anything for the FX5200 (40$) I decided to go for it as it was a DX9 compatible card and all other cards would have given me a bigger price difference (and I wasn't interested in high end cards as I didn't see any need for such an investment currently). You also may have purchased the 64-bit data-path variant of the FX5200 (non-Ultra).............same memory capacity (128Mbytes or 64Mbytes) but less-expensive to manufacture.......... 64-bit data-path == memory bandwidth 3.2Gbytes/sec 128-bit data-path == memory bandwidth 6.4GBytes/sec Quite a few FX5200 (non-Ultra) manufacturers omit the memory bandwidth on their specs and I challenge you to see that figure on the retail box. Deliberate ?? Do not confuse 128-bit processing with 128-bit data-path. If you do not see a clearly-stated memory bandwith spec. on a manufacturer's data sheet, watch out.......... Well it's an ASUS V9520/Magic/T 128MB card. I don't have the box around so I can't tell wether it's 64 bit or 128bit or if that is sai anywhere at all But I don't think my MX440 was a 128 bit card. Oh and if it helps then I'm running it at AGP 4x as mu mobo doesn't support AGP 8x. Well-informed as to desired performance and actual video card specs before purchase makes for greater happiness after purchase............ Well I read about the FX5200 and it seemed to satisfy all my requirements and I knew that it was a budget card, but hey MX440 was also a budget card so I didn't expect it to be worse than MX440 at some aspects. I still consider that I might have something a bit wrong as my 3dmark03 score is 860 without clocking and with some tuning I cat run at 940. At the same time I see people having with quite the same configuration scores from 800 - 1500. I can't quite get how the difference is so big? I saw that some of the high end scores were with AGP 8x so I guessed that this might be the issue, but I also found a score of 1300 with quite the same settings and conf I had except he had memory clock at 405MHz and I have at 333MHz (maximum by asus tweak utility is 360MHz so I can't quite understand how people get a lot more???) So I'm a bit disappointed in the performance compared to others. Might it be a problem with the card or is there a small thing I haven't noticed yet? I wouldn't be complaining if the card would be performing as it does for others as I know it's a budget card. But as it seems to be underperforming then I do tend to ask for help. Just for help here are my specs: OS: Win XP (without SP1) driver: have tried: 44.03, 44.67, 44.90 (current and with score 860) mobo: ASUS A7VK333 mem: kingston PC2700 (768MB) video: ASUS V9520/Magic/T (FX5200 128MB) my default gaming reso: 1280x1024x32 default benchmark reso: defaul (should be 1024x768x32 on 3dmark03) if you want any more details, then just ask. Mario |
#5
|
|||
|
|||
On Sat, 19 Jul 2003 12:44:31 +0300, Mario Kadastik
wrote: -Hello, - -John Lewis wrote: The FX5200 is a slow card. To get an exact comparison you need to prune the effects back to those available on the MX440. Use NVtweak or similar. where might I get NVtweak? - You probably should have considered a FX5200Ultra or better for your chosen screen-resolution and game software (and your FPS tolerance- threshold).. - Well actually the situation was that I had my MX440 and was quite happy with it (normal FPS in games and quite enough features to play games currently available. Also did play GTA Vice City and BF1942 and don't -remember having problems with smoke). - -But when it got fried around a month ago (I got periodic 10sek pauses while gaming and so every 30-60 sek) I decided not to replace it with the same card but to do an upgrade that wouldn't cost anything big. My only necessity was that the card should have tv-out (quite standard these days) and would be better than the one I had. As I practically didn't have to pay anything for the FX5200 (40$) I decided to go for it as it was a DX9 compatible card and all other cards would have given me a bigger price difference (and I wasn't interested in high end cards as I didn't see any need for such an investment currently). - You also may have purchased the 64-bit data-path variant of the FX5200 (non-Ultra).............same memory capacity (128Mbytes or 64Mbytes) but less-expensive to manufacture.......... 64-bit data-path == memory bandwidth 3.2Gbytes/sec 128-bit data-path == memory bandwidth 6.4GBytes/sec Quite a few FX5200 (non-Ultra) manufacturers omit the memory bandwidth on their specs and I challenge you to see that figure on the retail box. Deliberate ?? Do not confuse 128-bit processing with 128-bit a manufacturer's data sheet, watch out.......... Well it's an ASUS V9520/Magic/T 128MB card. I don't have the box around so I can't tell wether it's 64 bit or 128bit or if that is sai anywhere at all But I don't think my MX440 was a 128 bit card. - Oh and if it helps then I'm running it at AGP 4x as mu mobo doesn't support AGP 8x. Well-informed as to desired performance and actual video card specs before purchase makes for greater happiness after purchase............ - Well I read about the FX5200 and it seemed to satisfy all my requirements and I knew that it was a budget card, but hey MX440 was also a budget card so I didn't expect it to be worse than MX440 at some aspects. - -I still consider that I might have something a bit wrong as my 3dmark03 ..-score is 860 without clocking and with some tuning I cat run at 940. At -the same time I see people having with quite the same configuration -scores from 800 - 1500. I can't quite get how the difference is so big? -I saw that some of the high end scores were with AGP 8x so I guessed -that this might be the issue, but I also found a score of 1300 with -quite the same settings and conf I had except he had memory clock at -405MHz and I have at 333MHz (maximum by asus tweak utility is 360MHz so -I can't quite understand how people get a lot more???) - -So I'm a bit disappointed in the performance compared to others. Might -it be a problem with the card or is there a small thing I haven't -noticed yet? - -I wouldn't be complaining if the card would be performing as it does for .-others as I know it's a budget card. But as it seems to be .-underperforming then I do tend to ask for help. - .-Just for help here are my specs: - -OS: Win XP (without SP1) -driver: have tried: 44.03, 44.67, 44.90 (current and with score 860) -mobo: ASUS A7VK333 -mem: kingston PC2700 (768MB) -video: ASUS V9520/Magic/T (FX5200 128MB) -my default gaming reso: 1280x1024x32 -default benchmark reso: defaul (should be 1024x768x32 on 3dmark03) -if you want any more details, then just ask. - Mario http://www.oldi.ru/review/video/gffx/gffx.htm in other words your asus v9520 fx5200 magic t card is as 64bits version--- it as 2 bank of memory each side of the card... U shoulf replace that be a TD version witch is a 128 bits version and a lot more of feature like dvd support wire and game. Plus have seen thatit's memory chip was set to 400mghz and was build to operate at 500mghz. So i would say it a video card that will surely be overclocable. the difference betwen these card is about 10-20 $$$ But like u say if you satisfy with a 860 point wich (was about the score i had till a change) is to me the maximum of what you can get from it ... keep it. if you upgrade to the td version i'm sure u will get from it like 1000-1100 at 3dmark03 Plus notice that at future mark the fx5200 td and videosuite are not specifie, They are faster to, that's mean that it will have higher score. |
Thread Tools | |
Display Modes | |
|
|