|
Login to view full posting options
|
| Author |
Message |
Google
AdSense
|
|
| Back to top |
|
 |
powerarmour
--[papameister]--
Mod Squad
Joined: 13 Oct 2002
Posts: 11156
Location: Surrey, UK 'In Dead Space'
|
|
| Back to top |
|
|
powerarmour
--[papameister]--
Mod Squad
Joined: 13 Oct 2002
Posts: 11156
Location: Surrey, UK 'In Dead Space'
|
|
| Back to top |
|
|
powerarmour
--[papameister]--
Mod Squad
Joined: 13 Oct 2002
Posts: 11156
Location: Surrey, UK 'In Dead Space'
|
 Posted: Tue Dec 13, 2005 8:06 pm Post subject: |
|
|
|
IBM recently released a white paper showing real results on what the cell processor can really put out:
http://www-128.ibm.com/developerworks/power/library/pa-cellperf/?ca=drs-
At this year's E3 (or thereabouts) Sony proclaimed that their processor could achieve 200GFLOPS! However, according to IBM's white paper, only 155.5 GFLOPS was actually achieved (Table 4). BUT, IBM's tests used all 8 SPEs. The PS3 will only use 7 SPE's, due to manufacturing yield issues.
The efficiency of the Cell is 75.9% (Table 4), with of a theoretical peak of 201GFLOPs (Figure 5)--running 8 SPEs at 25.12GFLOPS apiece (Table 2). Similarly, the theoretical peak for the PS3's processor will be 176GFLOPS, using 7 SPEs at 25.12GFLOPS apiece. Assuming the same 75.9% effieciency, we could easily interpolate the PS3's Cell to be capable of 133.6GFLOPS.
The take home message is that with the PS3 being cabable of 133.6 GFLOPS and the Xbox 360 being capable of 115.2 GFLOPS, the PS3 is not nearly as far ahead of the Xbox 360 as we were lead to believe. All in all, we should expect relatively similar power coming from both consoles, processor power and ease of programming all considered. |
|
| Back to top |
|
|
powerarmour
--[papameister]--
Mod Squad
Joined: 13 Oct 2002
Posts: 11156
Location: Surrey, UK 'In Dead Space'
|
| Posted: Wed Mar 29, 2006 3:40 pm Post subject: |
|
|
|
More info:-
http://gametomorrow.com/blog/index.php/2006/03/24/cell-cant-texture/
| Quote: |
GameTomorrow
IBM leaders discuss the future of gaming
Cell Can’t Texture?
Blogged under Cell, consoles by Barry Minor on Friday 24 March 2006 at 12:14 pm
Much has been said about Cell’s presumed inability to texture map well. Given the small (256KB) local stores and DMA memory access, the SPEs were relegated by many to only handle nice streaming geometry type workloads. This seemed like an issue ripe for a little prototyping.
First, colleague Mark Nutter, implemented a software cache abstraction layer for the SPE giving us the ability to both hide the complexity of DMAs and benefit from transparent data reuse. Next, given the lessons learned from this paper, we tiled our textures, optimized our access patterns, and implemented several cache replacement policies. We then rewrote the shader in the Quaternion Julia Set Raytracer to add five cubemap texture lookup passes - 3 refraction lookups, a reflection lookup, plus a background lookup. These five texture lookups were then blended together with a fresnel calculation and modulated with the base lighting computation to form the final sample color.
The results were very pleasing.
Larger image: http://www.gametomorrow.com/minor/barry/frame.jpg
Quicktime H.264 movie (16MB so be patient)
We found that even with small 4-way set associative software cache sizes (8 KB), miss rates for this renderer were a low 7% and hit access times were only 12 SPE cycles.
Larger image: http://www.gametomorrow.com/minor/barry/graph.jpg
Using only seven 3.2 GHz SPEs we were able to raytrace 15 frames per second with a frame resolution of 1024×1024. The texture buffer held a cubemap with 1024×1024x16 bit texel faces resulting in a 12.5 MB texture buffer in XDR system memory. The performance penalty for using the five pass texture shader vs the lighting only shader was just 13%.
Our miss handler was implemented as a blocking function and we still have ideas pending to further reduce the 12 cycle software cache hit access time so we believe the 13% performance gap between the two shaders will continue to close.
|
*** *** ***
http://gametomorrow.com/blog/index.php/2005/11/30/gpus-vs-cell/
| Quote: |
GameTomorrow
IBM leaders discuss the future of gaming
GPUs vs Cell
Blogged under Cell by Barry Minor on Wednesday 30 November 2005 at 7:39 pm
Recently I came across a link on http://www.gpgpu.org that I found interesting. It described a method of ray-tracing quaternion Julia fractals using the floating point power in graphics processing units (GPUs). The author of the GPU code , Keenan Crane, stated that “This kind of algorithm is pretty much ideal for the GPU - extremely high arithmetic intensity and almost zero bandwidth usage”. I thought it would be interesting to port this Nvidia CG code to the Cell processor, using the public SDK, and see how it performs given that it was ideal for a GPU. First we directly translated the CG code line for line to C + SPE intrinsics. All the CG code structures and data types were maintained. Then we wrote a CG framework to execute this shader for Cell that included a backend image compression and network delivery layer for the finished images. To our surprise, well not really, we found that using only 7 SPEs for rendering a 3.2 GHz Cell chip could out run an Nvidia 7800 GT OC card at this task by about 30%. We reserved one SPE for the image compression and delivery task. Furthermore the way CG structures it SIMD computation is inefficient as it causes large percentages of the code to execute in scalar mode. This is due to the way they structure their vector data, AOS vs SOA. By converting this CG shader from AOS to SOA form, SIMD utilization was much higher which resulted in Cell out performing the Nvidia 7800 by a factor of 5 - 6x using only 7 SPEs for rendering. Given that the Nvidia 7800 GT is listed as having 313 GFLOPs of computational power and seven 3.2 GHz SPEs only have 179.2 GFLOPs this seems impossible but then again maybe we should start reading more white papers and less marketing hype.
|
| Quote: |
Comment by Luq Harith — December 1, 2005 @ 3:10 pm
I’ve been hearing about the PS3 Cell chip able to output 1080p graphics with HDR just by itself. Also, the E3 2005 PS3 Getaway demo is said to be rendered entirely by Cell without the help of a GPU. Can you clarify any of these news?
Comment by Barry Minor — December 2, 2005 @ 1:51 pm
Luq,
We have rendered real-time 1080p images with our Terrain Rendering Engine (TRE) using only Cell. All of our framebuffers are 128 bits per pixel. I don’t know anything about the “Getaway” demo so I couldn’t comment on how it was implemented.
|
*** *** ***
http://gametomorrow.com/blog/index.php/2005/07/26/beyond-polygons/
| Quote: |
GameTomorrow
IBM leaders discuss the future of gaming
Beyond Polygons
Blogged under Uncategorized, Cell by Barry Minor on Tuesday 26 July 2005 at 10:45 pm
First let me introduce myself. My name is Barry Minor and I have been on the Cell processor project since the fall of 2000. Before Cell I developed 3D graphics processors for IBM and Diamond under the FireGL brand.
Cell has been a great project and from the beginning we have focused the architecture around graphics and video processing. Once we had the architecture locked down I started writing a real-time ray-caster for Cell optimized around height-maps. As the design of the renderer progressed it became very apparent that Cell was not just good but stellar at such tasks. We found that we could ray-cast 720P images (1280×720) of complex scenes at frame rates greater than 30 frames/sec with a single Cell processors (50x a G5 VMX processor) and double that rate with a two way SMP configuration. Cell has the potential to move a new class of previously off-line rendering algorithms to real-time speeds thereby pushing us beyond polygon rasterization.
I think you will initially see hybrid approaches where backgrounds are rendered with ray-casting and foregrounds are rendered with GPU rasterized polygons but with the focus of people like Philipp Slusallek full blown real-time ray-tracing on Cell will be a reality.
|
| Quote: |
"To our surprise, well not really, we found that using only 7 SPEs for rendering a 3.2 GHz Cell chip could out run an Nvidia 7800 GT OC card at this task by about 30%."
"SIMD utilization was much higher which resulted in Cell out performing the Nvidia 7800 by a factor of 5 - 6x using only 7 SPEs for rendering. Given that the Nvidia 7800 GT is listed as having 313 GFLOPs of computational power and seven 3.2 GHz SPEs only have 179.2 GFLOPs this seems impossible but then again maybe we should start reading more white papers and less marketing hype."
"We have rendered real-time 1080p images with our Terrain Rendering Engine (TRE) using only Cell."
"We found that we could ray-cast 720P images (1280×720) of complex scenes at frame rates greater than 30 frames/sec with a single Cell processors." |
|
|
| Back to top |
|
|
Gilly
-- Meow --
Ultra nForced
Joined: 08 Feb 2004
Posts: 6016
Location: West Yorkshire | UK
|
| Posted: Wed Mar 29, 2006 6:07 pm Post subject: |
|
|
|
sounds powerful. makes me glad i like hardware as much as gaming  |
|
| Back to top |
|
|
daglesj
nForced˛
Joined: 10 Feb 2005
Posts: 1575
Location: on the cusp of uranus
|
| Posted: Wed Mar 29, 2006 8:18 pm Post subject: |
|
|
|
But will it do Busta Move?  |
|
| Back to top |
|
|
Tabajara
Mr. Bubbles
Ultra nForced
Joined: 30 Mar 2003
Posts: 7945
Location: I'd rather be surfing
|
| Posted: Tue Dec 19, 2006 1:50 am Post subject: |
|
|
|
| That could be Sony's initial plan, and it would really bring a whole new realism level... But somehow the hardware, and even more the software, wasn't up to it. Real time ray tracing games, that would put the 360 to shame |
|
| Back to top |
|
|
impar
PC Gamer
Moderator
Joined: 21 Mar 2003
Posts: 17806
Location: Portugal
|
|
| Back to top |
|
|
powerarmour
--[papameister]--
Mod Squad
Joined: 13 Oct 2002
Posts: 11156
Location: Surrey, UK 'In Dead Space'
|
| Posted: Tue Dec 19, 2006 9:28 am Post subject: |
|
|
|
| Tabajara wrote: |
| Real time ray tracing games, that would put the 360 to shame |
It depends, if Ray Tracing is good on essentially a multi-vector in-order processor like the Cell, it should be good on the Xenon too. |
|
| Back to top |
|
|
Tabajara
Mr. Bubbles
Ultra nForced
Joined: 30 Mar 2003
Posts: 7945
Location: I'd rather be surfing
|
| Posted: Tue Dec 19, 2006 2:09 pm Post subject: |
|
|
|
| powerarmour wrote: |
| Tabajara wrote: |
| Real time ray tracing games, that would put the 360 to shame |
It depends, if Ray Tracing is good on essentially a multi-vector in-order processor like the Cell, it should be good on the Xenon too. |
The Xenon doesn't have the specialized cores, they are the ones that would make a difference... And besides, my guess it that there would be one or more specialized a cells for the graphics, besides the one that the PS3 already has. |
|
| Back to top |
|
|
Google
AdSense
|
|
| Back to top |
|
|
|
|
nForcersHQ supporters: Impar, wardog, MiniMax, envoid, jasonh, Krips, VAIOMAN, Zetro, sensai, me&er, RElliott, ScottRempel, JHogarth, IcemanIX, Etraman, KA, Tabajara, Xiomberg, ruddywarrior, hookahmike, MMoore, RHodgson, RGilson, powerarmour, Sepal, RRavat, DigitalRuin, HammerheadTech, Bane, DHunt, Fat Jez, baikal, Jumpin' Jon, MGadzikowski, JBishop, RDeBok, JDavies, HDHyland, TMorris
|
|
| |
