Which is better, more RAM or a stronger GPU?

Well, in theory. In practice, that's a really nice way to find previously undetected race conditions. Besides, you would have to deploy a whole bunch of new devkits, QA would have to test on two platforms... fun.

But isn´t that still easier than testing a game for the PC?
 
But isn´t that still easier than testing a game for the PC?

QA's are more stringent on consoles (more time & money). On PC, devs can hammer out a quick patch without going through lengthy testing to see if the patch doesn't break anything else.
 
Well, in theory. In practice, that's a really nice way to find previously undetected race conditions. Besides, you would have to deploy a whole bunch of new devkits, QA would have to test on two platforms... fun.

Yeah this is all mostly just forum talk, I don't think any console maker in their right mind would introduce a better version of an existing box.


Crossbar said:
But isn´t that still easier than testing a game for the PC?

Console and PC are very different words. On console we have to be 100% sure that the game works. On PC they just lets the gamers be the Q/A department :)
 
My gut tells me the huge clock increase (50% really is massive). The RAM isn't optimal but still useable.

Depends on what area you might want to push.
 
Well, in theory. In practice, that's a really nice way to find previously undetected race conditions. Besides, you would have to deploy a whole bunch of new devkits, QA would have to test on two platforms... fun.

You wouldn't get a race condition between the GPU and the CPU if the former got a clock bump, or would you?
 
You wouldn't get a race condition between the GPU and the CPU if the former got a clock bump, or would you?

Why not? Synchronization works both ways. Contrived example: I'm on the PS3 and want to run over some rendertarget with the SPUs. I render into DDR, then DMA half of the buffer down into XDR (to save memory) and tell the RSX to wait for my SPUs. Once the SPUs are done, they release RSX, which copies down the second half of the buffer, overwriting the first half.
Now what happens if, say, I forgot to insert a barrier between the SPU computations and the RSX release. Usually, even with compiler reordering, the SPUs will be done long before RSX gets going.
But you just gave it a clock bump. ;)
 
Not unless your platform's a cellphone -- but does anyone do QA on those?

Developing for mobile phones in general can be a total nightmare, as it can run on one and not another, or even run on one specific model, but that tank on the same model in a different continent.
 
Why not? Synchronization works both ways. Contrived example: I'm on the PS3 and want to run over some rendertarget with the SPUs. I render into DDR, then DMA half of the buffer down into XDR (to save memory) and tell the RSX to wait for my SPUs. Once the SPUs are done, they release RSX, which copies down the second half of the buffer, overwriting the first half.
Now what happens if, say, I forgot to insert a barrier between the SPU computations and the RSX release. Usually, even with compiler reordering, the SPUs will be done long before RSX gets going.
But you just gave it a clock bump. ;)

Haha, ok I get it... no messing with console hardware! :)

I was wondering how much variability you would get from the console hardware itself? Like for example you don't know which SPE on each individual PS3 is disabled and theres a different latency going from certain SPEs. How does that effect your decisions since you only get one full bank of four SPEs and the other bank is unknown. Also beyond this, how much overall variation between different PS3s is there? Is it possible to have code which will break on one PS3 but not another?
 
Haha, ok I get it... no messing with console hardware! :)

I was wondering how much variability you would get from the console hardware itself? Like for example you don't know which SPE on each individual PS3 is disabled and theres a different latency going from certain SPEs. How does that effect your decisions since you only get one full bank of four SPEs and the other bank is unknown. Also beyond this, how much overall variation between different PS3s is there? Is it possible to have code which will break on one PS3 but not another?

Another potential difference that springs to mind is the speed of the harddrive as it can be swapped by the user. That could cause race condition hazards if some sloppy code is depending on slow read/write speeds. Do Sony test games with both slow/fast harddrive setups as well?
 
@Crossbar I think the PS3 may have been designed with faster HDDs in mind. Extremetechs testing didn't notice much in the way of a difference with a fast SSD installed IIRC. (Tired and about to go to bed)
 
I put a 7200 rpm hdd in my PS3 almost 6 months ago and have no problems with any games caching or running Burnout (or any PSN games). And the speed bump, while not a huge deal, is definitely noticeable and appreciated.
 
Back
Top