ok, lets try an example, you have a dual core, you load a game, you start playing it, one core has 100% load, the other has 0% load, tis will result in a 50% cpu usage, so with your explaination, this game is multithreade?
dont think so
Printable View
40% is more than 25%.. common logics tells us a QUAD core has 4 cores, therefore a core thats maxed at 100% only leads to 25% total CPU utilization.. So obviously 40% means more than 1 core is used.
my guess would be 3 tbh, but thats another topic
and wtf is that game that puts 100% on any single core jeeezuz...
That game is not multithreaded. But if you don't touch anything in CPU affinity etc. Windows will automatically balance the load, resulting in one core 80% and the other 20%, or whatever combination Windows wants to give you a final global utilization of 50%.
So what's your point? Any game that utilizes more than 25% in a quad is multithreaded. If you don't believe it, go grab some old game which is 100% CPU limited, or any other non multithreaded application like x264 encoder (old versions). You'll see constant 25% utilization, but not 100%-0%-0%-0%. 50%-50%-0%-0% or 25%-25%-25%-25% are two examples, but usually you'll see asimetric loading. But that doesn't change the fact that the application/game is single-threaded.
Have you ever done some serious SuperPi benching? Windows load balancing between cores is one your worst enemies ;)
Yes, it is multithreaded (or maybe not, keep reading). But don't be fooled by Windows task manager, because of the load of crap running in the background. In one of my previous examples (x264 encoding with old versions) in real life you'd see around 55-60% usage, that 5-10% caused by background processes. That example of yours (80-30% in a dual core) may be a single threaded game with crap in the background. In the Crysis situation you'll see higher %, way higher than any normal background program can use, so it's obvious that the game is using more than 1 thread. Maybe it uses only 2 or 3 threads, and maybe the game only needs 1% of threads 2&3 (so you can't "measure" them), but it IS multithreaded.
Crysis is not multithreaded?How come since it uses close to 90% my CPU's each core?And here is the utilization on a Quad Core.
90% i may be in a need for new glasses, but your core 0 shows a peak at 55% with an average of 40%, core1 average of 55%,core2 average of 30% and core3 average of 30%
looks like a poor implemented multithreading to me
http://www.gamespot.com/features/6182806/p-6.html
Dualcore faster then a quadcore at 1024 with medium settings
here you have it
http://www.tomsgames.com/us/2007/11/...ull/page2.html
http://www.extremetech.com/article2/...2209093,00.aspQuote:
Quad-Core Utilization and Overclocking the Video
Since I was working with a quad-core machine I wanted to make sure that the game was utilizing all four cores. I did some performance monitor captures while running the CPU benchmark to see if each core was getting some action.
EDIT: When I originally published this article and looked at the graph of the processor performance capture my first impression was that Crysis was utilizing all four cores - which is true. Crysis is passing instructions to all four cores but not maximizing their ability to handle instructions concurrently. All four cores look busy in the graph but the reality of it is that they are basically operating as a single-core and taking turns. In it's current state Crysis does not take full advantage of multi-core processing.
http://common.ziffdavisinternet.com/...=193291,00.gif
if it looks like a apple , smells like and apple, tastes like an apple........
its better than nothing tbh... framerate stability i think they said it was useful for..
I rather like Power Struggle, especially on Mesa, but i dont get why people dont like it?
I hope they optimize this game to allow AA usage.
Perkam
theres no such thing as a "multithread feature" :o?? Whats a "multithread feature" supposed to be? A piece of software is either multi-threaded, or single-threaded...
And I wouldnt believe everything I read on tomshardware, there was a thread before on this forum, some guy made a dual vs quad cpu test on crysis, there was a 2 fps advantage for the quad if I remember right. True, its not much, but its something..
whatever tho, people believe what they choose to
If the game uses more than one core it is multi threaded. Simply because a higher clocked dual core can outperform a quad core does not mean it is not multi threaded.
IE, lets say the game has four different threads, each on one of the quad core's four cores at 2.4GHz, utilizing 35% of each core. That means each thread is being run to the fullest a 2.4GHz core can run it as it is not taking 100% of the core. Now lets say another computer has the game running on the same four threads, two on each one of a dual core's 3.2GHz cores, utilizing 70% (2x 35%) of each core. Since it is still not needing to utilize 100% of the core, each thread is being run to the fullest at 3.2GHz. Continuing this example would leave a single core needing 140% utilization. Because that is impossible then a dual core will outperform it.
a multithreaded feature is like doing physics on another core, the game itself is not multithreaded, but like someone said above people are tend to believe what they think is correct
i dont ever care, i dont play crysis,crysis is just badly coded, it looks awesome and thats it, it doesnt even scale well with multiple GPU's, but again, like i care, believe what you want to believe
if you would have clicked on a link i provided you would ze a 2.66 dualcore outperforming a 2.4 quadcore
Seems we have different definitions, I consider any part of the code of one game to be part of the game itself :). For me physics IS a part of the engine, if a part of the engine runs on another core, then its multithreaded :)
It was me :D ;) I still encourage it ^_^
I'm out of here :)
I'm temted to sig you, but it's too large :D
Keep in your particular world with your "partially multithreaded" games/apps and your "multithreaded features".
That why I said it's useless for you. You are the type of guy that wouldn't see the real thing even if it's put right in front of your face.
It's the other way around mate ;)Quote:
no, all the evidence proved my point....
PS: I have another good reading for you by Jack: http://www.xtremesystems.org/forums/...&postcount=295