- This topic has 25 replies, 9 voices, and was last updated 16 years, 12 months ago by
.
-
Posts
-
@Lensman wrote:
Mugs – It might be worth checking that you’ve got the full AGP bus speed. If your card can do 8x, for example, but your motherboard only supports 4x then that will throttle the transfer of data between main memory and card and kill your benchmark. Or maybe your motherboard can do 8x, but for some reason has been set to a lower speed.
Poke around the display information dialogs to dig this out, and maybe your BIOS settings too.
Worth a check anyway, just to rule it out.
Thanks for the idea Lensman.
My motherboard is an Asus P4P800 with the Intel 965PE chipset which supports AGP 8x. I have checked the bios and not sure what to set my agp apature size to 256MB as I had it on 128MB for my old card which had 128MB ram on it. my new one has 256MB ram.
There is also another option about fast write which i have always had on and the ati smartgart thing says 8x.
My old Radeon 9800 ran at full speed quite happily also for the last 4 years or so.
may be its the drivers at fault. I have installed the ones that came with the card now. but need to do more testing tonight to see if that improves things.
3451 in 3d mark 06 is a very reasonable score for the card you have and compares well the tomshardware benchmarks of similar cards
Well I suppose the older ram and single core cpu will be holding it back a little bit.
I am thinking that the 3d mark 2003 test i did was not acurate for my new card.
basically the image quality was set for performance and AA and AF always defaulted to app decideds with my old radeon and the GF6800GT when testing those.
Having looked at it now with my new gfx card 4xAA and 8xAF is on and app decides is not ticked. also image quality is set for quality. So this alone I would suspect could reduce the score a fair bit as the new card is having to do more work than the other 2?
still got more testing to do with it. It was just that in some games its running a bit low on the fps than i was expecting that is all, namely my racing game (rFactor) i had that running at 1280×1024 with 4xAA 8xAF and max details and shadows and with a grid of 20 cars sat in front of me the fps is around 16fps which was a bit low, probably find its the shadows killing the performance a bit, but it doesnt half look fantasic if i can just get the FPS increased!
I have also installed the drivers off the cd that came with the gfx card now so need to see how they work compared to the 7.10 ones
Ok it seems that my graphics card is fine and it was because the catalyst drivers did different defaults (I installed the complete lot with the control centre this time)
I went through it all and changed the aa and af and so on to app decides rather than having it permantely on.
ran the 3d mark 03 test and I get 14954 3d marks instead of 9,996 3d marks like before. so that against the 4,867 3d marks I got with my old radeon 9800 makes this new card look like it might be quite good performance wise!
Hmm well this is interesting. Theres a guy who bought the same agp card as me but with an AMD athlonXP 3000 and reckons he had COD4 running at 1600×1200 with 4xAA and 8xAF at 60fps lol.
I will have to try this, as I tried that but with 1024×768 and it didnt feel like 60fps lol.
Out of interest do all you guys play yours games with AA and AF? As it always seems to be the case that all gfx cards are crippled by it, this is 3 cards in a row that struggle with AA and AF. always played games with it off in the past.
Mind turned these up to max to play BF1942 and it doesnt half make a difference to things in the distance.
I also noticed that AF does not seem to hit performance very much? its the AA that kills it? is this a correct assumption?
Some games incorparate it more than others, but with CoD you do notice a difference with the AA on, i personally run all games with full AA etc
Heh well I just tried running cod4 at 1600×1200 with no aa or af and it looks quite nice and fps seems to be great. will try with aa on in a min. although I bet it will kill the fps. It runs ok at 1024×768 with 4xaa and 8xaf ok I noticed.
Right well it looks like it may be my old single core P4 that is the bottleneck in my system.
Found a well rated ASROCK motherboard on ebay for £25 to use as a stop gap with my current components and stick in a nice Core2Duo cpu (Core2Duo E4500) which on all the charts i have seen is much faster than my old P4. (P4 2.8ghz 800fsb 512kb cache socket 478 Northwood)
Also going to get 2 sticks of 1GB Cosair DDRII 667 ram to replace the 1.5GB of Cosair DDR333 ram i have now (this should bring a performance increase surely?)
This board will take my AGP X1950XTX ok and hopefully let it run to its max speed. Should cost me about £100 to do this and in the future for upgrades I can get a PCI-E gfx card and new motherboard only, which is a lot cheaper than all in one go
Been having a browse online at 3d marks ORB section and many Core2Duo people are getting much higher 3d mark scores with those chips and the same card as me.
So looking at around 7500 3d marks with a core2duo cpu at around 2.5ghz or so and ddrII ram. I ran the test again on my current P4 with the card overclocked and got 3750 3d marks and 3700 without being overclocked.
Thats up on my previous score of 3451 before.
Those are the overall scores. Mine gets 1780 in SM2.0 and 2167 SM3.0/HDR the charts on toms hardware show 2035 SM2.0 for the X1950XTX so cpu might help things hopefully.
Read another article on toms hardware with radeon x1950’s agp’s on an old athlon xp (like my p4) and then they compared the same ones with a much newer athlon dual core with ddr II ram and the fps increased by quite a margin as the cpu was not bottlenecking it lol.
I am going to get a Core2Duo E2180 2ghz cpu and 2GB DDR II 5400 ram. should cost around £85 + the £25 for the m/b I am getting off ebay. Lots of people rate this cpu quite highly and its cheap 🙂 alternative would be an E4500 which has 2MB caches instead of 1MB.
- You must be logged in to reply to this topic.