Posted on 03/30/2004 2:11:05 AM PST by Swordmaker
First - Load the Google Toolbar, by visiting Google and loading the toolbar from tools. On the current homepage, click on more and load.
The Toolbar gives you an excellent popup blocker while also enhancing your websearches with a highlight feature for the word(s) you are searching.
Go back to the Google tools page and click on Google Labs, (boiling test-tube)
then click on Google Compute.
A default name will be assigned by Google. However, should one of us, enter a new name eg: freepers4Google, we could all form a team that could go to the top very soon.
I installed this tool on about 20 PCs last year, and now sit in the top 10 with nearly 700 assignments completed.
You have the option at install, or any time later to use the distributing computing command either at standard or conservative mode, and then another option at 100%, 80% lower.
Even at 100%, standard, the google process only works when the user or other processes are active. It is set at the very lowest priority for your cpu. You will not notice it affecting your PC, unless you watch your cpu cycles, and then will notice your PC is humming along at 100% all the time.
I have it installed on my wife's three PCs and she has never even noticed it is there, and she is extremely sensitive about the speed and response from her PCs.
Just to suggest that someone here could take the lead, using the google tool, establish a team name, provide a link for assistance, then occasionally update all of us as to how the team is doing.
If you leave your PC on most of the time, it is a great opportunity to use the electricity effectively.
On a 3 ghz cpu the average assignment take about 2 days, while a celeron 500 mhz might take 7-10 days. when completed, you will have a file of about 1 megabyte to upload to stanford. This works even with dial-ups, as the only connection required is between assignments, to receive the next project and to upload your results.
The Mac OSX team, made up of people on Apple/PPC machines, has 1300 members, and has completed 188,604 units at the moment. The HardOCP team, made up of people on Intel/AMD boxes, has more than 5,000 members and has completed 1,802,804 units. Now, let's see - if I want to do some distributed computing, and have to choose between one platform or the other, which one should I target? Hmmmmm. That's a tough one. I guess I could choose the platform with the greater total computing power, but then again, the other one does have a fruit on the box...
For distributed/parallel computing, the speed of any one node is not an appropriate measure of performance - the only thing that really matters is aggregate performance. That's not what Apple wants anyone to take from this, but then again, they've always been better at marketing than just about anything else. ;)
By the reading, it appears that they still have a lot of slow G4s in the group. Given the PPC 970's vastly superior floating point performance (to both Intel and G4), likely 50% clock boost by the end of the year (with corresponding bus speed boost), and 10-20% speed increases due to the new compiler, the G5s should be really pumping up the OS X numbers through this year.
Superior performance which, thus far, remains officially undocumented. Where's the SPEC results, if it's so hot? Even Apple appears to have pulled those fakey-SPEC rate results from their website now.
From my own look at it, it appears to me that a single G5 is approximately on a par with a single P4, on a clock-for-clock basis, and so far I've seen nothing to disabuse me of that notion.
Anyway, for distributed computing - particularly distributed computing on someone else's computers, where I don't have to worry about how much they cost - individual node speed really doesn't matter, only aggregate performance, like I said. Given that there are five times as many HardOCP members as there are OSX members, the average PPC box will have to be five times faster than the average x86 box for Team OSX to match the performance of Team HardOCP. And that's just not happening. Even if the PPC team triples the work-unit-per-node performance of x86, which is wildly unrealistic, the fact that there are so many more x86 boxes still puts them way ahead. Team OSX may not fall behind as rapidly with new compilers and the like, but I expect that they'll still be falling behind, if for no other reason than because they're simply outnumbered.
Now, if we're shopping for a supercomputer, that might be a different story, but that's not the example we're presented with here.
I'm waiting for SPEC too. Looking at third-party PPC/Intel results shows a dual G5 2 GHz roughly equal to a dual Xeon 3 GHz, less in integer and better in FP. No way is a PPC 970 1 GHz as slow as a P4/Xeon 1 GHz. The processor architecture is just much more advanced, among other things capable of keeping 215 instructions in flight at a time.
The Xeon/P4 has a big problem with bus, as shown recently in an Opteron/Xeon shootout, where the Intel's slow, shared bus can't keep the processors fed. Plug 20% faster Xeons in the current system, and you will barely get more speed. Intel is slapping lots of L3 cache on the chips to alleviate the problem, but the end cost of the chips is enormous. The G5's bus runs at half the speed of the processor, and each processor gets its own bus, so the G5 will scale well to 3 GHz this fall, with a 1. GHz bus to keep those processors fed. Intel seems to be hitting a wall.
Given that there are five times as many HardOCP members as there are OSX members, the average PPC box will have to be five times faster than the average x86 box for Team OSX to match the performance of Team HardOCP.
That's why distributed computing projects make mutltiple versions.
I haven't been following it for the last few months or so, so perhaps it's changed - got a link?
Anyway, the important thing, from Apple's perspective, is that it's competitive - you don't have to hold your nose when buying one ;)
I still think that Intel is making a mistake by not pushing I2 harder, but I guess we'll have to see where they go with x86-64. The latest batch of P4s/Xeons have an 800 MHz frontside bus, and so should the first x86-64s, which ought to keep them in the game, at least. It'll be fun to watch, as always...
Here's one. Apple does a lot better with MP because it's based on a processor designed from the ground up to work in MP environments. And it has a lot of room to grow. All this not so much thanks to Apple really, but mainly to IBM.
I still think that Intel is making a mistake by not pushing I2 harder, but I guess we'll have to see where they go with x86-64
The I2 is just way too expensive. It competes better with the POWER processors, the G5's big daddy. Compatibility with current apps is also a problem. It has an x86 compatibility mode, but it is very slow and doesn't have the SSE instructions.
The latest batch of P4s/Xeons have an 800 MHz frontside bus, and so should the first x86-64s, which ought to keep them in the game, at least.
The problem is that it's a shared bus. The Opteron starts getting an edge at 2x MP, and starts blowing away the higher clocked Xeon at 4x and especially 8x as the Xeon can't get fed data fast enough. This is a problem the G4 had. G5s, with their independent busses, will have the same advantage over Xeons.
I'm hoping Intel gets its act together, or AMD/Apple might get lazy.
The I2 is just way too expensive. It competes better with the POWER processors, the G5's big daddy.
That's because Intel is playing its usual twenty-year-old game of "it's a SERVER chip. Got it? SERVERS only". If/when they decided to mass-market it, you'd see the price drop quick enough, just like it has with every other "it's a SERVER chip" that Intel has ever made ;)
Compatibility with current apps is also a problem. It has an x86 compatibility mode, but it is very slow and doesn't have the SSE instructions.
Actually, I2 does SSE, but not SSE2. Anyway, the idea was to ditch IA-32/x86 altogether, remember?
The problem is that it's a shared bus. The Opteron starts getting an edge at 2x MP, and starts blowing away the higher clocked Xeon at 4x and especially 8x as the Xeon can't get fed data fast enough. This is a problem the G4 had. G5s, with their independent busses, will have the same advantage over Xeons.
So far that's not translating into much of a handicap versus AMD or G5. Eventually they'll have to rearchitect the frontside, but eventually everyone will have to scrap everything they're currently doing - in the here and now, it's not a serious problem.
Anyway, what do I care? We can always switch to AMD, right?
Errr, well, one of us can ;)
Already did. All my new computers are AMD, but with a switch to Mac at the next systems. I just love that OS, and have great faith in IBM to keep that chip advancing, especially since it ties in with their POWER server line.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.