amount of ram in the consoles
This topic is locked from further discussion.
[QUOTE="jakjbt"]not to sure but this amzes me i play oblivion on my "512MB RAM" 360 for six hours...no slowdown then i play oblivion on my "1GB RAM" (dual core processor)vista and after maybe 2 hours the gameplay framrate gets really choppy and slowdown is very bad....why????
MetroidPrimePwn
You set the system settings too high for your computer to handle smoothly, maybe? Oh, and also, the problem may be that Vista is a freaking system hog...
1GB is not enough for Vista...vista alone uses almost that amount
from a developer standpoint, ram differences between the two consoles is huge.
developers struggle with the ps3 because it has a set amount of ram (256 which isnt really alot) for assets like static meshes, light maps, shadow maps, materials and shaders, textures. Thus you cannot go over the allotted amount of ram for graphics, same goes for physics, navmesh and AI on the other 256.
xbox360 (and pc) work the exact same way where your total amount of memory will adapt and accomodate what is needed at the time of render, thus it is easy for the xbox and pc to balance out the load. The problem with the ps3 is that if you bust (go over your memory budget) in either field that ive listed above, you are screwed, you will get lag and possible system crash, thus you have to optimize your game.
This is why devs have more ease developping for the xbox360 and PC.
That is why when you have multiplats, the same game that comes out on xbox will run better on it (will get better fps) whereas the ps3 will struggle. This is because developpers do not want to over-optimize their game for looks sake (and the reputation of the ps3) so they will take the FPS hit over having to make the game look ugly (lower texture resolution, geometry polycount, dumb the AI down, use less physics and interactive objects, etc).
I'm not hating on the ps3, i actually sold my xbox360 to get a ps3, and I am enjoying it (and my PC) alot.
Sorry if this comes as offensive to PS3-husbands.
[QUOTE="ChiefMahwoo"]not much the ps3 has two 256mb ramsthe xbox 360 has 512mb ram while the ps3 has 2 256mb ram. I have a ps3 but just wondering how much difference this makes with the systems.
micky4889
256MB XDR Main RAM @3.2GHz -CPU
256MB GDDR3 VRAM @700MHz-GPU
360 has
512 MB GDDR3 RAM @ 700 MHz DDR (SHARED)
Unified architecture
What do you notice?
ps3 256 mb ram CPU ram operates at 3.2 GHz.
360's 512mb shared ram operates at 700 MHz.
This is an example of how ps3 creates more with little. Clearly the ps3's 256mb CPU Ram operates way faster than the xbox360
.Even if 360's Ram in unified.
xbox360 shares its ram.
This slows down 360's CPU. Like intergrated G-cards do with Computers. Intergrated G-cards share the computers ram, i know i have one. This slows down my computer alote.
i have heard that the 256 xdr at 3.2ghz ram would be equivelent to the speed of 400 mb of GDDR3 at 700mhz because of the extra speed
LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
[QUOTE="WARxSnake"]from a developer standpoint, ram differences between the two consoles is huge.
developers struggle with the ps3 because it has a set amount of ram (256 which isnt really alot) for assets like static meshes, light maps, shadow maps, materials and shaders, textures. Thus you cannot go over the allotted amount of ram for graphics, same goes for physics, navmesh and AI on the other 256.
xbox360 (and pc) work the exact same way where your total amount of memory will adapt and accomodate what is needed at the time of render, thus it is easy for the xbox and pc to balance out the load. The problem with the ps3 is that if you bust (go over your memory budget) in either field that ive listed above, you are screwed, you will get lag and possible system crash, thus you have to optimize your game.
This is why devs have more ease developping for the xbox360 and PC.
That is why when you have multiplats, the same game that comes out on xbox will run better on it (will get better fps) whereas the ps3 will struggle. This is because developpers do not want to over-optimize their game for looks sake (and the reputation of the ps3) so they will take the FPS hit over having to make the game look ugly (lower texture resolution, geometry polycount, dumb the AI down, use less physics and interactive objects, etc).
I'm not hating on the ps3, i actually sold my xbox360 to get a ps3, and I am enjoying it (and my PC) alot.
Sorry if this comes as offensive to PS3-husbands.
Jacobistheman
a pc's ram is split just llke the ps3's it is split between the video and the main system but there is alot more so it is less problems. DO SOME RESEARCH before you let more stupid out onto the page
The OS on your PC OBVIOUSLY needs RAM as well as all the background programs. However this does not mean that the RAM your OS uses cannot be used for the game because it doesnt rely on specifc instructions to bank only on video or physics or any other asset.
on 32bit systems where your max is 3GB of ram, 2GB total will be dedicated to the game and those 2GB are not split into banking assets or physics, they bank everything. on 64bit systems, you can allocate all the ram you want for your game. I'm talking solely about the ram the game uses and nothing else. Read carefully before acting stupid and calling others stupid
LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
cool_guy2000
That "different way" your talking about takes up an entire SPU and still has bandwidth limitations.
The term you are looking for is "TurboCache". The RSX is a TurboCache GPU, it is capable of using system RAM for additional VRAM.LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
cool_guy2000
[QUOTE="cool_guy2000"]The term you are looking for is "TurboCache". The RSX is a TurboCache GPU, it is capable of using system RAM for additional VRAM.Yeah, but that's just remapping addresses; we've had that sort of thing since 1998. Try going in the other direction - as I understand it, there's a way to read from VRAM into system memory, but it's too slow to be useful.LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
skektek
[QUOTE="skektek"][QUOTE="cool_guy2000"]The term you are looking for is "TurboCache". The RSX is a TurboCache GPU, it is capable of using system RAM for additional VRAM.Yeah, but that's just remapping addresses; we've had that sort of thing since 1998. Try going in the other direction - as I understand it, there's a way to read from VRAM into system memory, but it's too slow to be useful.LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
lowe0
http://research.scea.com/ps3_deferred_shading.pdf
"The pixel shader runs on the Synergistic Processing Elements (SPEs) of the Cell/B.E. and works concurrently with the GPU to
render images. The system's unified memory architecture allows
the Cell/B.E. and GPU to exchange data through shared textures.
The SPEs use the Cell/B.E. DMA list capability to gather
irregular fine-grained fragments of texture data generated by the
GPU. They return resultant shadow textures the same way. The
shading computation ran at up to 85 Hz at HDTV 720p
resolution on 5 SPEs and generated 30.72 gigaops of
performance. This is comparable to the performance of the
algorithm running on a state of the art high end GPU. These
results indicate that the Cell/B.E. can effectively enhance the
throughput of a GPU in this hybrid system by alleviating the
pixel shading bottleneck."
It is unquestionable that this process is useful. Moreover, you weren't able to illustrate or describe what it is accurately.
[QUOTE="skektek"][QUOTE="cool_guy2000"]The term you are looking for is "TurboCache". The RSX is a TurboCache GPU, it is capable of using system RAM for additional VRAM.Yeah, but that's just remapping addresses; we've had that sort of thing since 1998. Try going in the other direction - as I understand it, there's a way to read from VRAM into system memory, but it's too slow to be useful. The Cell's reading from VRAM is very slow (~16MB/s) but there is no reason why the CPU would need to read from VRAM.LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
lowe0
[QUOTE="lowe0"][QUOTE="skektek"][QUOTE="cool_guy2000"]The term you are looking for is "TurboCache". The RSX is a TurboCache GPU, it is capable of using system RAM for additional VRAM.Yeah, but that's just remapping addresses; we've had that sort of thing since 1998. Try going in the other direction - as I understand it, there's a way to read from VRAM into system memory, but it's too slow to be useful. The Cell's reading from VRAM is very slow (~16MB/s) but there is no reason why the CPU would need to read from VRAM.LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
skektek
It's called deferred shading, and on 5 SPUs, it generates 30.72 gigaops of performance, apparently.
[QUOTE="skektek"][QUOTE="lowe0"][QUOTE="skektek"][QUOTE="cool_guy2000"]The term you are looking for is "TurboCache". The RSX is a TurboCache GPU, it is capable of using system RAM for additional VRAM.Yeah, but that's just remapping addresses; we've had that sort of thing since 1998. Try going in the other direction - as I understand it, there's a way to read from VRAM into system memory, but it's too slow to be useful. The Cell's reading from VRAM is very slow (~16MB/s) but there is no reason why the CPU would need to read from VRAM.LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
BobHipJames
It's called deferred shading, and on 5 SPUs, it generates 30.72 gigaops of performance, apparently.
Yes. I read your link. Concurrent textures are stored in the XDR bank where both the Cell and the RSX have fast read/write access (as opposed to using the VRAM where the Cell has very slow read throughput).[QUOTE="Jacobistheman"][QUOTE="WARxSnake"]from a developer standpoint, ram differences between the two consoles is huge.
developers struggle with the ps3 because it has a set amount of ram (256 which isnt really alot) for assets like static meshes, light maps, shadow maps, materials and shaders, textures. Thus you cannot go over the allotted amount of ram for graphics, same goes for physics, navmesh and AI on the other 256.
xbox360 (and pc) work the exact same way where your total amount of memory will adapt and accomodate what is needed at the time of render, thus it is easy for the xbox and pc to balance out the load. The problem with the ps3 is that if you bust (go over your memory budget) in either field that ive listed above, you are screwed, you will get lag and possible system crash, thus you have to optimize your game.
This is why devs have more ease developping for the xbox360 and PC.
That is why when you have multiplats, the same game that comes out on xbox will run better on it (will get better fps) whereas the ps3 will struggle. This is because developpers do not want to over-optimize their game for looks sake (and the reputation of the ps3) so they will take the FPS hit over having to make the game look ugly (lower texture resolution, geometry polycount, dumb the AI down, use less physics and interactive objects, etc).
I'm not hating on the ps3, i actually sold my xbox360 to get a ps3, and I am enjoying it (and my PC) alot.
Sorry if this comes as offensive to PS3-husbands.
WARxSnake
a pc's ram is split just llke the ps3's it is split between the video and the main system but there is alot more so it is less problems. DO SOME RESEARCH before you let more stupid out onto the page
The OS on your PC OBVIOUSLY needs RAM as well as all the background programs. However this does not mean that the RAM your OS uses cannot be used for the game because it doesnt rely on specifc instructions to bank only on video or physics or any other asset.
on 32bit systems where your max is 3GB of ram, 2GB total will be dedicated to the game and those 2GB are not split into banking assets or physics, they bank everything. on 64bit systems, you can allocate all the ram you want for your game. I'm talking solely about the ram the game uses and nothing else. Read carefully before acting stupid and calling others stupid
what you are saying is that the pc has combined memory for the graphics card and the main system which is WRONG. The pc has Graphic memory built into the card and memory for the main system processes unlike the 360 which combines the two. I would think before replying a second time you would check what you are saying.
There really isn't much dfference, except that the video memory and programs share that 512MB pool. This usually allows the programmer to allocate memory however they want to. The PS3, I believe, has 256MB dedicated to both video memory and programs/system. I am hearing that the video memory can be shared with the system, but I'm not an expert at these things.
The games can be optimized for the system, because the programmers know what they are working with. Thus making it preform well (in some cases).
This is old news though...
Haven't a majority of developers already stated that PS3 is the more powerful console? I know they have stated that the peak power is much harder to reach because of development difficulties with the system architecture.
Each console has access to 512MB of RAM they access it in different ways. 360 has an easier time because of the unified architecture but PS3 has access to it's memory as well, extra code must be written to do so though. I do feel that 360 is the best out of the two when it comes to multiplatform games because 360's architecture is just more developer friendly.
PS3 is again like the PS2 it isn't developer friendly but does have an amazing amount of power. The Cell is a processing powerhouse and Uncharted showcased that quite well.Haven't a majority of developers already stated that PS3 is the more powerful console? I know they have stated that the peak power is much harder to reach because of development difficulties with the system architecture.
Each console has access to 512MB of RAM they access it in different ways. 360 has an easier time because of the unified architecture but PS3 has access to it's memory as well, extra code must be written to do so though. I do feel that 360 is the best out of the two when it comes to multiplatform games because 360's architecture is just more developer friendly.
PS3 is again like the PS2 it isn't developer friendly but does have an amazing amount of power. The Cell is a processing powerhouse and Uncharted showcased that quite well. I_Helios_I
I'm sure the dev's are right, that the system is more powerful, but it can only be as powerful if the components can handle it? My idea is if they really want to get that power to work, you gotta compensate by putting the correct amount of hardware. If this Cell processor is so powerful combined with the video card, why limit with the amount of memory and put more in there... I believe this would break the programming boundries of the system, allowing rich games to be developed with more ease.
My opinion of course,
[QUOTE="lowe0"][QUOTE="skektek"][QUOTE="cool_guy2000"]The term you are looking for is "TurboCache". The RSX is a TurboCache GPU, it is capable of using system RAM for additional VRAM.Yeah, but that's just remapping addresses; we've had that sort of thing since 1998. Try going in the other direction - as I understand it, there's a way to read from VRAM into system memory, but it's too slow to be useful.LOL at the clueless people thinking the PS3 only has access to 256 ram....the PS3 is capable of using 512(add it up, geez) but just done in a different way to that of the more easier 512 the 360 has access too.
get informed.
BobHipJames
http://research.scea.com/ps3_deferred_shading.pdf
"The pixel shader runs on the Synergistic Processing Elements (SPEs) of the Cell/B.E. and works concurrently with the GPU to
render images. The system's unified memory architecture allows
the Cell/B.E. and GPU to exchange data through shared textures.
The SPEs use the Cell/B.E. DMA list capability to gather
irregular fine-grained fragments of texture data generated by the
GPU. They return resultant shadow textures the same way. The
shading computation ran at up to 85 Hz at HDTV 720p
resolution on 5 SPEs and generated 30.72 gigaops of
performance. This is comparable to the performance of the
algorithm running on a state of the art high end GPU. These
results indicate that the Cell/B.E. can effectively enhance the
throughput of a GPU in this hybrid system by alleviating the
pixel shading bottleneck."
It is unquestionable that this process is useful. Moreover, you weren't able to illustrate or describe what it is accurately.
What you linked and what we were talking about aren't the same thing. What you're describing looks like a post-rendering pass back to the XDR (which is slow, but would only have to be done once), followed by more intensive work on the SPEs (which have a fast link to the XDR) and a final write to the GDDR3.Also, if it chews up 5 SPEs, that's probably pretty academic. You'd be competing for processing resources with everyone else - AI, physics, you name it.
[QUOTE="WARxSnake"][QUOTE="Jacobistheman"][QUOTE="WARxSnake"]from a developer standpoint, ram differences between the two consoles is huge.
developers struggle with the ps3 because it has a set amount of ram (256 which isnt really alot) for assets like static meshes, light maps, shadow maps, materials and shaders, textures. Thus you cannot go over the allotted amount of ram for graphics, same goes for physics, navmesh and AI on the other 256.
xbox360 (and pc) work the exact same way where your total amount of memory will adapt and accomodate what is needed at the time of render, thus it is easy for the xbox and pc to balance out the load. The problem with the ps3 is that if you bust (go over your memory budget) in either field that ive listed above, you are screwed, you will get lag and possible system crash, thus you have to optimize your game.
This is why devs have more ease developping for the xbox360 and PC.
That is why when you have multiplats, the same game that comes out on xbox will run better on it (will get better fps) whereas the ps3 will struggle. This is because developpers do not want to over-optimize their game for looks sake (and the reputation of the ps3) so they will take the FPS hit over having to make the game look ugly (lower texture resolution, geometry polycount, dumb the AI down, use less physics and interactive objects, etc).
I'm not hating on the ps3, i actually sold my xbox360 to get a ps3, and I am enjoying it (and my PC) alot.
Sorry if this comes as offensive to PS3-husbands.
Jacobistheman
a pc's ram is split just llke the ps3's it is split between the video and the main system but there is alot more so it is less problems. DO SOME RESEARCH before you let more stupid out onto the page
The OS on your PC OBVIOUSLY needs RAM as well as all the background programs. However this does not mean that the RAM your OS uses cannot be used for the game because it doesnt rely on specifc instructions to bank only on video or physics or any other asset.
on 32bit systems where your max is 3GB of ram, 2GB total will be dedicated to the game and those 2GB are not split into banking assets or physics, they bank everything. on 64bit systems, you can allocate all the ram you want for your game. I'm talking solely about the ram the game uses and nothing else. Read carefully before acting stupid and calling others stupid
what you are saying is that the pc has combined memory for the graphics card and the main system which is WRONG. The pc has Graphic memory built into the card and memory for the main system processes unlike the 360 which combines the two. I would think before replying a second time you would check what you are saying.
I'm not talking about VRAM when mentionning the PC, because the system RAM helps in games as well. I'm talking about the 2GB of system memory, apart from the 768MB of VRAM in a 8800gtx for example
What you linked and what we were talking about aren't the same thing. What you're describing looks like a post-rendering pass back to the XDR (which is slow, but would only have to be done once), followed by more intensive work on the SPEs (which have a fast link to the XDR) and a final write to the GDDR3.
Also, if it chews up 5 SPEs, that's probably pretty academic. You'd be competing for processing resources with everyone else - AI, physics, you name it.
lowe0
You do realize this is a CPU? Your evaluation has nothing to do with what I was attempting to do....which is to prove that it is obviously not a useless procedure.
Think a Core Duo could pull those benchmarks with both general purpose cores/
Yes, I realize that the entire design of the Cell focuses on highly parallel, single-precision floating-point math on vectors. Gee, I wonder what other type of chip is designed to do that?You do realize this is a CPU? Your evaluation has nothing to do with what I was attempting to do....which is to prove that it is obviously not a useless procedure.
Think a Core Duo could pull those benchmarks with both general purpose cores/
BobHipJames
I also realize that the work is being done back on the CPU. Do you think, just maybe, the part where I point out the work being transferred back to the SPEs might have indicated that?
[QUOTE="BobHipJames"]Yes, I realize that the entire design of the Cell focuses on highly parallel, single-precision floating-point math on vectors. Gee, I wonder what other type of chip is designed to do that?You do realize this is a CPU? Your evaluation has nothing to do with what I was attempting to do....which is to prove that it is obviously not a useless procedure.
Think a Core Duo could pull those benchmarks with both general purpose cores/
lowe0
I also realize that the work is being done back on the CPU. Do you think, just maybe, the part where I point out the work being transferred back to the SPEs might have indicated that?
My point was that the Cell processor's performance in this particular application isn't going to be a limiting factor in real-world applications....it's not as though the typical videogame is going to require that 5 SPEs be utilized exclusively for graphics processing while the GPU stands idle.
Even in this demonstration, the GPU is rendering the app concurrently, although I did not link and probably cannot find (assuming they exist) the independent benchmarks for RSX's performance.
Further, my point was that in this IDEAL, "academic" environment, the Cell processor was clearly not inhibited by sharing textures between the two different kinds of memory.
"We have explored the potential of the Cell/B.E. for
accelerating graphical operations in the PLAYSTATION®3
computer entertainment system. This system combines the
Cell/B.E. with a state of the art GPU in a unified memory
architecture. In this architecture both devices share access to
system memory and to graphics memory. As a result they can
share data and processing tasks."
You clearly didn't understand how the operation was being conducted. It says quite clearly that both the GPU and the CPU utilize BOTH pools of memory. "Share access." "Both devices." Then it enumerates specifically both SRAM and VRAM, I.E. XDR and GDDR3. Your assumption was wrong. This wasn't a quick pass through to the XDR (which is "slow?"), this was both VRAM and SRAM.
"The processors are
connected to each other and to system memory through a
high speed Element Interconnect Bus (EIB). This bus is
also connected to an interface (IOIF) to the GPU and
graphics memory. This interface translates memory
accesses in both directions, allowing the PPE and SPEs
access to graphics memory and providing the GPU with
access to system memory. This feature makes the system a
unified memory architecture since graphics memory and
system memory both are visible to all processors within a
single 64-bit address space."
"We used 32 bit
float RGBA textures for all data. The textures for these
attachments may be allocated in linear, swizzled or tiled formats in either GPU or system memory. We
experimented with all combinations of texture format and
location in order to find the combination that gave the best
performance.
GPU performance is highest rendering to native tiled
format in GPU memory. The performance advantage is
high enough that it is worth rendering in tiled format and
then reformatting the data to linear allocation for processing
by the Cell/B.E. In order to minimize the latencies incurred
by the SPEs in accessing this data we reformat the data into
system memory rather than GPU memory."
"Tables 1 and 2 show that the shading calculation can be
sped up to meet any realistic performance requirement.
The monochromatic shader ran at 85 Hz using 5 SPEs and
at 34 Hz using 2 SPEs. Videogames are typically rendered
at 30 or 60 frames per second. Shading calculations should
generally run at these rates, but for shadow generation it is
possible to use lower frame rates without affecting image
quality. It would also be possible to use shadows generated
at 720p resolution with a base image rendered at a higher
1080p resolution (1920x1080 pixels).
Table 3 analyzes the time spent waiting for DMA
transactions to complete. This was as much as 27% of the
total time. Note that if we were able to remove all of this
DMA waiting the performance on 5 SPEs would reach 116
frames per second as indicated by the"no waiting'' data in
table 1."
Um....lawls? Doesn't this put your statement right in the basket alone? They achieved 85 Hz even with DMA waiting....without DMA waiting, they would reach 116 FPS on the SPEs. This is a 27% reduction....is that really as HOBBLING as you made it out to be? I mean, seriously.
"We can also conclude that the performance of the Cell/B.E.
is superior to a current state of the art high end GPU in that
we achieved comparable performance despite performance
limitations and despite using only part of the available
processing power. Our current implementation loses
substantial performance due to DMA waiting. This results
from the fine-grained irregular access to memory and is
specific to the type of shaders we have chosen to
implement. We have explored shaders based on shadow
mapping [15] which require evaluating GPU fragments
generated from multiple viewpoints. These multiple
viewpoints are related to each other by a linear viewing
transformation. Gathering the data from these multiple
viewpoints requires fine-grained irregular memory access.
Deferred Pixel Shading on the PLAYSTATION®3
8
This represents worst-case behavior for any memory
system."
Their recommendation? Possibly, this: "Removing the
computation from the GPU effectively increases the frame
rate, or more likely, the geometric complexity of the models
that can be rendered in real time."
They state that the GPU can utilize system memory although it's best left to VRAM until it transfers data (this is how I understood it) at which point it's tiled for system memory and transferred to the SPEs....this sounds like an application for VRAM and SRAM transferrence to me. The VRAM can be used by the GPU and the unified memory architecture can be used to transfer information, textures, shading information...the latencies incurred in some rendering methods don't seem to stifle performance to the point where it'd be totally without benefit to use concurrent rendering, information sharing.
Please Log In to post.
Log in to comment