Does anyone know if there will be a new version optimized to take advantage of the Cell processor's SPEs or at least the power processor (AltiVec) element?

Haven't heard anything about it yet, but I'd second the motion for such a client ;)
The PS3 is going to run on Linux, so it should be possible.

Since we're on the subject of Sony, they have previously done fundraising to fight certain diseases in their online game EverQuest 2. They had an enemy spawn at an easily accessable place and every time someone killed him they'd donate a small amount - added up to a few thousand dollars. They worked with Kurt Schilling (Baseball Player) on it... I wonder if they'd consider mentioning rosetta@home? The game has VERY high system requirements, so a lot of unused high end computers just waiting to hear about Rosetta ;)

---

This post on the thread that Einstein@home hosts said that:

The Cell architecture differs quite much from the current standard CPU designs, so while potentially being more powerful, much more effort is needed on the software side.

If you leave aside the synergistic processing elements (SPEs), you get a rather ordinary PPC-based machine. Given the projected price of the PS3, you'd be better off with a bunch of Macs than PS3s.

To make use of the SPEs, thus accesing the full power of the Cells, you don't only need a new compiler; like e.g. with processing data in the Graphics Controller you have to implement explicit, software-controlled memory accesses / tranfers. It will take some man-months of work to implement this.

Thus for Einstein@Home this would make sense for clusters of many hundreds of identical machines, exclusively running Einstein@Home, but if it's for a few dozens of users with one or two machines each, the time would be better spent by improving the Windows App by, say, 1%.

(http://einstein.phys.uwm.edu/forum_thread.php?id=3644)

The applications that are applied for the processor are yet to appear, especially on PlayStation3.
Please read the thread for more information.

---

IBM is also going to be releasing blades which can cluster and use the Cell processor: http://www-03.ibm.com/press/us/en/pressrelease/19198.wss

Imagine running BOINC on one of these if it was optimized to take advantage of it!

Related with the IBM blade server, there appeared an article which comments around the Cell processor.

The expectation and topic to Cell in the blade server (written in Japanese)

I found it with the search engine just now, so there may be of course better article, but this also seems worth for you to read. The Japanese article may be hard for non-native. Try translate through Google or Altavista. But even through them it'll stay hard to understand:( You should search another article if you don't understand it at all.

I'll explain the point of the article.

The problem which Cell would face could be, according to the writer, so large amount of cost, money, (and hours) would be required to make applications which make the best use of the processor. It's sure that the use for science like BOINC projects wouldn't always be influenced by costs. But even the ability of it would be highest, it might not spread, if it would cost greatly us to make applications. Therefore, to spread the Cell, the keypoint is whether environment where we develop and use applications would be consolidate or not.

Spreading of a thing over the market is important if it's expected to be used in general. For the makers, whether the processor will spread or not is actually the most important issue. It could vanish from the market if it wouldn't be sold well. It would be a problem for consumers, too. The processor couldn't be supported if it desappered from the market. Without support or plenty of information it must be hard for developers to make the software which is applied for the architecture of the CPU.

I think, unless the environment is consolidate, BOINC projects developers wouldn't be willing to make applications which supports the Cell, since:

The Cell architecture differs quite much from the current standard CPU designs,

therefore,

if it's for a few dozens of users with one or two machines each, the time would be better spent by improving the Windows App by, say, 1%.

Here we can agree with the importance of the processors spread:)

---

I think, unless the environment is consolidate, BOINC projects developers wouldn't be willing to make applications which supports the Cell, since:

The Cell architecture differs quite much from the current standard CPU designs,

therefore,

if it's for a few dozens of users with one or two machines each, the time would be better spent by improving the Windows App by, say, 1%.

Here we can agree with the importance of the processors spread:)

I admit that I may be taking this a bit out of context, but I want to address both the "inner" quoted points.

The difference in the cell architecture is huge. I'm not 100% convinced we can make effective use of it for Rosetta, because the SPU's (where the cell gets its huge Flops count) only have a very small amount of memory: 256K for both code and data. At least that's the specs for the cell going into the PS3. They may make a bigger memory area for the SPU's for other systems, but it will never grow above L2 cache sizes. 512K maybe, 1Meg absolute max.

The question then becomes "Can Rosetta run for timeframes measured in milliseconds only touching L2 cache?" That's the sort of memory access pattern needed to get good performance out of the SPU's, because the moment you need to go "outside" the 256K arena, you need to use DMA to flush data to main memory, and to pull in the new data. As it stands, I don't think we're anywhere near that.

That said, addressing the second point, I do think it would be worth at least taking a look. The PS3 uses the Cell, and all indications are that you'll be able to run Linux on the PS3. Sony may not approve, but given past history people will find a way. And you can reasonably expect that there will be a fairly large base of installed PS3's in a fairly big hurry after the PS3 hits the shelves.

Indeed, it might be worth trying to contact Sony, and see if they'd be interested in this. It might be a big selling point for them: "Use your PS3 to cure diseases when you're not playing games on it."

---

It might be a big selling point for them: "Use your PS3 to cure diseases when you're not playing games on it."

This will be attractive to PS3's consumers:)

It were good if we could test PS3 for BOINC computing. Anyway I'll purchase it as soon as possible, not to play games but testing it!

---

PS3 uses the Cell, and all indications are that you'll be able to run Linux on the PS3. Sony may not approve, but given past history people will find a way.

Linux is going to be pre-installed on the PS3. I'm not sure if the idea of coding for cell processors has been given enough thought by the people over at Einstein.

Console gamers often tend to be competitive - perfect for the credit system. With the PS3 expected to sell about 5 million consoles in the first 6 months that's a decent sized market. And if the technology ends up in home computers in the future, the work will have to be done anyway.

Sony could actually use it as a way to support their "IT'S A COMPUTER!" statement - having it considered a computer instead of a console is a HUGE deal for them, because it means they have to pay less taxes in the EU. They lost a case claiming the PS2 was a computer, cost them hundreds of millions.

---

By the way,

Sony could actually use it as a way to support their "IT'S A COMPUTER!" statement - having it considered a computer instead of a console is a HUGE deal for them, because it means they have to pay less taxes in the EU. They lost a case claiming the PS2 was a computer, cost them hundreds of millions.

Actually SCE's CEO said that PS3 was the computer, not just the game console. But because Nintendo, also a Japanese game maker, said Revolution(Wii) as the game console completely, PS3 is also being regarded as the same kind of thing. It's said that therefore SCE determined to install linux on PS3 to be seen it as a computer as we mean always.

By the way,

Sony could actually use it as a way to support their "IT'S A COMPUTER!" statement - having it considered a computer instead of a console is a HUGE deal for them, because it means they have to pay less taxes in the EU. They lost a case claiming the PS2 was a computer, cost them hundreds of millions.

Actually SCE's CEO said that PS3 was the computer, not just the game console. But because Nintendo, also a Japanese game maker, said Revolution(Wii) as the game console completely, PS3 is also being regarded as the same kind of thing. It's said that therefore SCE determined to install linux on PS3 to be seen it as a computer as we mean always.

Hope I understood you right:

Back when the PS2 released, they also tried to get it classified as a computer - just 2 weeks ago the UK's highest court refused their appeal: (yes, they fought this for almost 6 years)

Dismissing Sony's entire argument as a joke, Lord Justicee Chadwick said Sony's "skeleton argument" went "beyond what can be regarded as acceptable written advocacy." He continued:

I am not protesting about [the Sony appeal's] inordinate length, nor about its discursive quality, nor about its frequent and unnecessary resort to hyperbole, although all those unappealing features are present...

My concern is with the repeated aspersions that are cast in that document on the intellectual honesty of the High Court Judge from whose decision this appeal is brought.

So now they HAVE to do everything in their power to get the PS3 declared a computer. That's why every time someone calls it a console now, they'll be corrected that it is a PC. There are no different "models" but different "configurations" etc. It's a legal thing they have to win this time, they already take a big loss on the initial sales as it is.


It likely sounds stupid to fight so hard for the distinction, but I believe the console import tax for the EU is another 10% or so. Fights over such small things are more common than we might think: Those in the US will likely also start seeing ads or other notifications (or already have them?) not to use "google" as a verb - ie you can't "google" something, you "look something up on google". Because if "googling" becomes a commonly used verb, Google loses their trademark for it. Xerox successfuly fought that btw (at least they're delaying what may be happening unavoidably), as photocopying was often called "xeroxing" even if it wasn't on a Xerox machine. Aspirin (now commonly used as something against headaches) didn't - nor did escalator (used to be a brand name). Very long list of such cases ;)

---

PS3 uses the Cell, and all indications are that you'll be able to run Linux on the PS3. Sony may not approve, but given past history people will find a way.

Linux is going to be pre-installed on the PS3. I'm not sure if the idea of coding for cell processors has been given enough thought by the people over at Einstein.

Console gamers often tend to be competitive - perfect for the credit system. With the PS3 expected to sell about 5 million consoles in the first 6 months that's a decent sized market. And if the technology ends up in home computers in the future, the work will have to be done anyway.

Sony could actually use it as a way to support their "IT'S A COMPUTER!" statement - having it considered a computer instead of a console is a HUGE deal for them, because it means they have to pay less taxes in the EU. They lost a case claiming the PS2 was a computer, cost them hundreds of millions.

Linux Preinstalled? You learn something new every day. :)

I can't help but make the observation that it'd be an incredibly smart move for someone to try to see what it'd take to port Rosetta to the cell. It might be possible, it might be close to impossible. I've done any amount of "inner loop" work on the VU's in the PS/2 and from what I've been told, the SPU's are a whole lot like VU's on major steroids. The big question becomes whether we can find a way to apply all that CPU horsepower to Rosetta.

After a longish discussion elsewhere on these forums about the applicability of SSE, let's just say that I'm not 100% convinced that it'll be a perfect fit to try to run Rosetta on the Cell (or any SIMD CPU for that matter). But we sure as hell need to try.

---

hello,

If you'd like to know around Cell processor more, check official information on SCEI, Cell Broadband Engine. There are plenty of documents, enough for us to understand what the Cell is like. It seems worthwhile to read and study them for those who are going to write or optimise applications on the CPU.

---