![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
-
Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.
What would be the bottleneck
- Thread starter Kid Cruel
- Start date
mlambert890
Limp Gawd
- Joined
- Jan 8, 2007
- Messages
- 175
This is a more complex topic than a simple "bottleneck/not bottleneck" answer. I started into this on the sticky thread re: 8800GTX game testing, but you have to think of the 3D pipeline.
Part of the process is handled by the CPU, part (most these days) is then managed by the GPU. In addition, the CPU obviously runs game AI, tracks objects, runs the OS, etc. A faster CPU completes its work faster allowing for a quicker hand-off to the GPU. A faster GPU gets the scene from start point to display that much quicker. Imagine it in terms of the CPU "sketching" the scene and the GPU then coloring it in and displaying it. The CPU may be able to "sketch" the scene at, say, 100fps. In this case 100fps would be the ceiling. The GPU may be able to render the scene at 300fps, but it can never exceed the 100fps ceiling set by the CPU. As you up the res and the effects, the GPUs job gets harder and its ceiling comes down. When analyzing component integration, you have to consider where is the CPUs ceiling at a given complexity vs. what is the GPUs capability.
At a certain resolution level, even the 8800GTX SLI runs out of steam. At lower resolutions, I suspect that only the most aggressively overclocked Core 2 will hit the point where the CPU is theoretically capable of delivering higher framerates than the X1950 could render (ie - CPU can sketch the scene at a higher framerate than the GPU could ever hope to render it). You can demonstrate this easily by the fact that framerates at a res like 1280x1024 continue to scale up even on an X1950 as you go from E6300 to X6800 to overclocked C2 monster. What I would like to do is set up a test where a given card (X1950 would be a good one) is tested with a range of procs just like that to see if/where the point is that the CPU starts waiting on the GPU. No time/energy for that though and none of these dedicated sights seem willing to do that. They just fuzzy it up and talk about "bottlenecks" and "needing" big CPUs with big GPUs. Massive oversimplification.
Its all about the overall performance profile of the system based on the peak performance of each subsystem.
Part of the process is handled by the CPU, part (most these days) is then managed by the GPU. In addition, the CPU obviously runs game AI, tracks objects, runs the OS, etc. A faster CPU completes its work faster allowing for a quicker hand-off to the GPU. A faster GPU gets the scene from start point to display that much quicker. Imagine it in terms of the CPU "sketching" the scene and the GPU then coloring it in and displaying it. The CPU may be able to "sketch" the scene at, say, 100fps. In this case 100fps would be the ceiling. The GPU may be able to render the scene at 300fps, but it can never exceed the 100fps ceiling set by the CPU. As you up the res and the effects, the GPUs job gets harder and its ceiling comes down. When analyzing component integration, you have to consider where is the CPUs ceiling at a given complexity vs. what is the GPUs capability.
At a certain resolution level, even the 8800GTX SLI runs out of steam. At lower resolutions, I suspect that only the most aggressively overclocked Core 2 will hit the point where the CPU is theoretically capable of delivering higher framerates than the X1950 could render (ie - CPU can sketch the scene at a higher framerate than the GPU could ever hope to render it). You can demonstrate this easily by the fact that framerates at a res like 1280x1024 continue to scale up even on an X1950 as you go from E6300 to X6800 to overclocked C2 monster. What I would like to do is set up a test where a given card (X1950 would be a good one) is tested with a range of procs just like that to see if/where the point is that the CPU starts waiting on the GPU. No time/energy for that though and none of these dedicated sights seem willing to do that. They just fuzzy it up and talk about "bottlenecks" and "needing" big CPUs with big GPUs. Massive oversimplification.
Its all about the overall performance profile of the system based on the peak performance of each subsystem.
