Jump to content

bioduken

Members
  • Content Count

    62
  • Joined

  • Last visited

About bioduken

  • Rank
    Member
  1. Hello again! I have been in some discussions about gaming performance. Have also read game code (how it is done) and think that I now have the real explanation how it works. I first thought (before reading about gaming and development there) that games takes advantages of higher resolutions. This is done in normal applications. Games don’t seem to do this. The resolution does not influence how the game behaves in relation to the work processor needs to do. The processor will do the same work on 640x480 compared to 2560x1600. What I said before that was because I thought games did send more information to the gpu and needed to calculate more information dependent of the resolution. This isn't the case. The reason why AMD did perform better on some tests here ( http://www.overclockersclub.com/reviews/intel_q9450/ ) is probably because higher MIN FPS. Intel has one big L2 cache. This big L2 cache is very helpful in order to get higher average FPS and max fps. When the video card isn’t that big bottleneck, if you run on one simple road etc the gpu don’t need to work as much. Then Intel will boost FPS a lot and those FPS gained there will add to total FPS and will also increase the average FPS. If you increase the resolution the workload for the GPU is increased. On higher resolutions the video card will cut all those high max values or high FPS values that Intel gets when there is little to do in the game. This will also increase the influence of the lowest FPS on total score. If one game is running at 800x600 and there isn’t anything that happens. All that is done is some running one simple road. This could make the FPS go very high on Intel, lets say 150 FPS and AMD has 100 FPS. When there is some action, fights and a lot of enemies the workload for the processor will increase and the cache isn’t as advantageous as before and the cpu will probably also increase the data sent to the video card. If AMD then will have higher FPS (50 in this sample) and Intel gets 40 but this fight is for one small period, then Intel gets higher average FPS. This action section in the game will also be more dependent of the processor. Now the resolution is increased to 1920x1200 and MAX FPS is down to 55 (because of the video card) when there is little action. That means that the 150 FPS max score is cut and there isn’t that much that will hurt average FPS for the Intel and its large L2 cache. When there is action now in the game at 1920x1200 Intel might get 20 FPS and AMD gets 25, then this higher MIN FPS will make more to average FPS on the total score. If you want to test which processor that is best for one specific game then you should run the game at lowest possible resolution but have high detail and other settings that can be high. Let say 640x480. Then try to pick parts that are very heavy for the processor and check the lowest FPS values for each processor. The processor that has the highest low areas is the best processor for that game.
  2. If you say anything good about AMD Phenom you are a troll?
  3. Performance list showing how processors scale for the game Race Driver Grid LIST - Processors (reversed order)
  4. Intel Q9450 vs Phenom 9850 GPU = ATI HD3870 X2 Game: Call of Duty 4 Settings: 1920x1200, 4xAA 16xAF 4xAA 16xAF, everything MAX Q9450 = 65 Phenom 9850 = 69 Game: Crysis Settings: 1280x1024, 1xAA/1xAF, DX9 everything HIGH Q9450 = Min FPS 22.46, Max FPS 61.71, Avg 41.96 Phenom 9850 = Min FPS 29.82, Max FPS 63.20, Avg 47.89 Game: STALKER Settings: 1920x1200, 16xAF everything MAX Q9450 = 83 Phenom 9850 = 84 Game: TDU Settings: 1920x1200, 4xAA/16xAF, everything HIGH Q9450 = 45 Phenom 9850 = 47 Game: Half Life 2 Settings: 1920x1200, 4xAA/16xAF, MAX everything. Q9450 = 295 Phenom 9850 = 300
  5. I am just waiting for that test to arrive I just hope it isn
  6. English isn't main main language . I am sitting in Europe up in the north. Also I have dyslexia and another reason doing this to improve my language skills (or lack of) in writing english
  7. Have you bought eggs and tomatoes I know I am right because I understand reading about hardware (been in this business for many years). If you isn
  8. I am working and reading forums between compiling (software developer) Don
  9. It does appear in every game, it just hit the roof on CoH. It can’t go any higher and the difference is so small that it is almost even. If you check the others you will see that is more difference between the cards when they are alone compared to using them in crossfire. If they had compared with Phenom I am sure that it should perform better
  10. more proof http://news.firingsquad.com/hardware/ati_r...mance/page9.asp We
  11. Well, read about Nehalem, if you understand hardware you will see that they focus A LOT on low latency and the FSB is gone. They have in fact smaller L2 caches compared to Phenom in order to reduce latency and make the L2 cache a bit faster. Why do you think that the engineers do this? Is it a just for fun. Remember, one of the easiest things to do is staying stupid.
  12. If something does 100 fps and another do 50 fps. Which is faster? That
  13. yes I do! A couple of years ago i was developing a CRM system and we used SQL Server as a database. This system (I and two others where programmers there) won some test and the company was selling more SQL Servers than any other company here. Microsoft loved us. But I also know how "programs" work. All these partner programs and how you need to create software in order to sell with often differ a lot for real use. Sometimes programmers think, why the hell do I develop this when nobody will use it. The reason is “selling”. We where doing a lot of functionality in order to impress the right people. As a partner you are often forced to do various things, Microsoft was going to get the browser market (this was when Netscape had that market). We were forced to install IE if we should use some functions in the operating system. The thing is that I am the only one that is presenting facts, there hasn’t been others here that has facts that that tells me wrong. Some has presented facts about the cpu and how fast that is but that’s another issue. They are mixing up different technologies. Some thinks that servers are a different world, that there are totally different applications there. It isn’t. I am doing threaded server software (software for analyzing information stored in databases) now so I am very well aware how it works. I could go deeper what the real problem is. It’s about latency, bandwidth is in fact not the actual problem but that is a much more complicated discussion. To get a fast computer you don’t want latency. When you clock the FSB latency is decreased. Less latency and hardware that are doing the work can focus on that and don’t need to wait. AMD has done a lot for decreasing latency but they market it as more bandwidth. The reason for that is that few understand latency and how it effects speed. Sometimes people say that the feel that the computer responds faster. If you clock one intel and notice that is faster that’s because less latency.
  14. Why do you get angry? Do you have some knowledge about marketing?
  15. ok, fair enough. I have tried to find test that contradict the FSB problem but I cannot find it. If I find something then it just confirms it. I have found some older reviews but they have been done with older cards (PCI Express 1.0) and the FSB should be able to handle that speed if game is single threaded or mainly single threaded. As I have said I don
×
×
  • Create New...