[PC] Grand Theft Auto IV
- Thread starter eezing
- Start date
There's no announcement of anything but a 360 release, but when GTA4 released Rockstar was also denying the possibility of a PC port.
Okay, I installed my new GTX 285, changed everything to max (textures, everything) except view distance (set to 21) to keep inside the resources (what happened to Rockstar saying you needed 1.5 GB VRAM to have high textures?) and ran the benchmark a few times.
First time: 54.03 avg
Set the view distance to 41, the game allowed me to do this (I have all restrictions enabled) and benchmarked:
Second time: 47.74 avg
This is fine, I thought, but just to experiment, I reduced the view distance to 30
Third time: 46.99 avg
Well, lower performance? Interesting. Reduced the view distance back to 21, ran benchmark
Fourth time: 45.43 avg
Stupid memory leak. In disgust, I quit the game. I am going to email Rockstar support. I hope they are aware of this, but in case they are not, well, they'll be now.
I need a new patch!
First time: 54.03 avg
Set the view distance to 41, the game allowed me to do this (I have all restrictions enabled) and benchmarked:
Second time: 47.74 avg
This is fine, I thought, but just to experiment, I reduced the view distance to 30
Third time: 46.99 avg
Well, lower performance? Interesting. Reduced the view distance back to 21, ran benchmark
Fourth time: 45.43 avg
Stupid memory leak. In disgust, I quit the game. I am going to email Rockstar support. I hope they are aware of this, but in case they are not, well, they'll be now.
I need a new patch!
Yes, I am assuming it is a video memory leak, but as I said, I sent an email to Rockstar support with a DXdiag.txt and the four benchmark txts. Let's see what they say.
Forgot to say that the fact that the game starts popping in textures and objects really really close to me sooner the higher amount of VRAM I have enabled the game to use. I can run medium textures with everything else to max and after a little while, it's as if I have the noprecache commandline argument enabled.
Same feeling i got from GTA4. It has pretty graphics and physics, but the rest is just a basic shell of a city, much like we have seen in the previous generation GTA's.
New update 1.0.3.0 today. The link they posted isn't updated as of this writing but the patch is live on GFWL.
I'm still in dialogue with Rockstar support. I'm still working to get through to them about my performance issues. Here's how I provoke it: run the benchmark 15 times or so without closing the game between each benchmark run. Run the benchmarks from in-game, one session. I see my average FPS drop by 10 FPS and the game turns all choppy and stuttering. If I play the game after running the benchmark, I get behaviour that looks an awful lot like if the -noprecache command line argument is specified, that is, objects and things disappear if I turn the camera away from it. I drove down to the park and set my car in the shadows of a tree, I roated the camera so that the tree was behind me and the shadows disappeared. I turned the camera back and the tree reappeared after a little while. Anyway, I lose 20% performance after about 20-30 minutes of play or 12-15 runs of the benchmark.
I'm still in dialogue with Rockstar support. I'm still working to get through to them about my performance issues. Here's how I provoke it: run the benchmark 15 times or so without closing the game between each benchmark run. Run the benchmarks from in-game, one session. I see my average FPS drop by 10 FPS and the game turns all choppy and stuttering. If I play the game after running the benchmark, I get behaviour that looks an awful lot like if the -noprecache command line argument is specified, that is, objects and things disappear if I turn the camera away from it. I drove down to the park and set my car in the shadows of a tree, I roated the camera so that the tree was behind me and the shadows disappeared. I turned the camera back and the tree reappeared after a little while. Anyway, I lose 20% performance after about 20-30 minutes of play or 12-15 runs of the benchmark.
http://www.gtaforums.com/index.php?showtopic=403690 said:
Absolutely.
I have to admit, my initial negative reaction to this game now seems a little unjustified.
Ok, it still performs worse that you would expect for a console port but its one hell of a game, and Rockstar are certainly putting the effort in to please the PC community.
Good on them!
Can anyone else try what I mentioned above? Run 13 or 15 or whatever benchmark runs in succession without closing the game between each run. I lose about 20% from the first benchmark to the thirteenth or so.
Use the highest settings you can get away with without using any commandline arguments that remove restrictions.
My settings are: everything at absolute max except textures which are set to medium. Remember that texture filtering have a step above very high which is called highest. If the settings are playable or not is what this is about, it is the delta in average FPS between the first and the last runs.
Post your CPU, GPU, GPU driver and OS too.
i7 920@3.2 GHz, GTX 285, GeForce 182.08 WHQL, XP32SP3
Use the highest settings you can get away with without using any commandline arguments that remove restrictions.
My settings are: everything at absolute max except textures which are set to medium. Remember that texture filtering have a step above very high which is called highest. If the settings are playable or not is what this is about, it is the delta in average FPS between the first and the last runs.
Post your CPU, GPU, GPU driver and OS too.
i7 920@3.2 GHz, GTX 285, GeForce 182.08 WHQL, XP32SP3
Is the bar for PC games that low? Except for the control stuff, everything in that list is a bug fix. 'Gosh, thank you Rockstar for fixing the broken stuff in the game we bought. You're cool folks.'
Though maybe you should be positive, they never fixed the texturing bug on the 360, as far as I can tell.
About the comments in earlier pages about performance increase being little next to nothing and arguments relating to clock frequency vs FSB bandwidth.
It crossed my head when watching some benchmarks on the i7 cores that the reason behind such small increases in performance on dual cores and quad cores with either higher clock or FSB is not (mainly) due to bad coding, but more likely inter-core bandwidth from the dual/quad cores (not the FSB) as the i7's have much higher inter-core bandwidth from what i could understand.
It crossed my head when watching some benchmarks on the i7 cores that the reason behind such small increases in performance on dual cores and quad cores with either higher clock or FSB is not (mainly) due to bad coding, but more likely inter-core bandwidth from the dual/quad cores (not the FSB) as the i7's have much higher inter-core bandwidth from what i could understand.
I don't get this part, higher clock increases perfomance linearly unless the GPU perfomance is the bottleneck as seen in all proper CPU benchmarks for GTAIV. Higher FSB should also improve perfomance (when exceding VRAM amount) AFAIK since when the VRAM is not enough the higher bandwidth will increase the write/read speed to RAM. Improving the speed of the VRAM<->RAM texture write/read operations.
Still i7 is faster (inter-core connection, cache, improved tech, etc?) but increased clock does improve perfomance as expected.
EDIT: I benched and checked recorded results in SS 2009 Professional regarding inter-conenction bandwidth and latency.
My CPU E8400 - 10.7GB/sec, 37ns
Ext QX6950 ----- 20.3GB/sec, 23ns (Core 2 extreme 3.3GHz QX6950 1333MHz FSB)
QX6950 ---------- 18GB/sec, 29ns (1333MHz FSB)
Q6600 ----------- 14.9GB/sec, 35ns
Phenom 9750 -- 3.4GB/sec, 137ns
Phenom 8750 -- 2.9GB/sec, 159ns
Core i7 920 ----- 33.8GB/sec, 17ns
Athlon FX60 ----- 2.9GB/sec, 95ns
Last edited by a moderator:
Trying again.
Next time I get a spare weekend mate, that damn benchmark takes ages to load up after each run (unless you have a batch file that could be used?)
Similar threads
