Ok, I've never optimized and so here is the question:

I downloaded the CPU-Z and it said I have an "Intel Pentium III EB".


Every optimized version I see says it is for SSE, Athlon or another. I've never seen "EB" in the optimized downloads I've seen. So, as Paul says in his wicki.....get it right!

I'm running Windows ME.

Any help in pointing me in the right direction would be appreciated.

Please don't just give me a site to go to, I've been to many. I'm looking for the EB version to download for my computer.

Edit> In the CPU-Z:

Code Name is Coppermine. Under Chipset: Intel i810E

Bill: There is no field for SSE...maybe because i do not have it?

mmciastro: YAOSCW-K-r8.1 says: Optimized with Intel's IPP 4.1 library for P3 and compatible Intel processors "with SSE plus AMD AthlonXP" so I'm not sure this is right either?

Thanks

---

[img][/img]

"EB" is meaningless. Most optimization is for some type of math ability on the chip; SSE, SSE2, SSE3, Altivec, etc. If you have none of these, there is likely no way to optimize using "chip specific" compile-time methods, only through actually changing the source code. Did CPU-Z show anything at all in the field where SSE etc. would be? Did you look at the Wiki instructions on optimizing for Windows? I know there is an optimized SETI application for Linux for P3, that uses caching (and wants 512MB RAM) but don't know of one for Windows... perhaps someone does and will give you a URL, but frankly there's just not much room with a P3 to do very much. The "stock" client may be your best bet.

my p3-500 went from 11:42 down to 8:30-9:25 using TMR's yaoscw K r8.1. available at the Marisan site, I'm using win98 first edition

I downloaded CPU-Z on to that P3. In the "processor" catagory and under "instructions" it lists MMX and SSE. This is the info you're looking for.

Ok, if I did it right with the optimizing....this is what I get for Bench Marks:

Old: 722 Whet
958 Dry

New: 764 Whet
1337 Dry

Is this good, is this bad, is it anything, did I do anything.....LOL

mmciastro: you said with a 500 yours are about 9 hours. I have a 933 and they are saying about 10 hours now....lol

Maybe I'll have to run a few and see how long they actually take?

---

[img][/img]

did you update the BOINC client or the seti application? the seti application does not do benchmarks

When I first booted that ole machine and loaded 5.2.2, seti estimated 24 hours but did them in 11:42. I installed the optimized application and crunching dropped to 8:30-9ish. My estimated has dropped to 11:10. I watched it drop from 24 to 23 to 22 to etc.

Ok, if I did it right with the optimizing....this is what I get for Bench Marks:

Old: 722 Whet
958 Dry

New: 764 Whet
1337 Dry

Is this good, is this bad, is it anything, did I do anything.....LOL

That seems within the normal range of variability; the same processor can give bigger differences than that, depending on other load. The optimized SETI app should have made no difference to the benchmarks. (After running it for a couple of weeks - minimum - if your claimed credit is way too low, you can look at optimized BOINC Clients, but the gain you'd see there is VERY minimal, and all on credit claim, not on crunch time.) You didn't say which one you decided to use - the yaoscw K r8.1 mm recommended for SSE? If it's working, then your P3 has SSE, otherwise you'd see a nice crash... Using SSE, your actual crunch times should be cut by about 25-30%. BOINC will gradually apply a correction factor to the estimated times for your results. The estimate is originally figured based on your benchmarks and the "stock" application; now that you're using an optimized application, in about 2-3 days, you'll see the estimate get lower, until it's pretty close to reality.

I don't know if my "25-30%" guess for SSE is accurate - please let us know! I know that Altivec optimization cut my times on a G4 from almost 8 hours down to around 2 hours. The newer the cpu, or the more difference from the 'default' used by UCB in compiling, the more optimization can help!

CAN YOU TELL ME WHAT MB AND CPU YOU HAVE?

---

Randy,

In the Wiki guide to using an optimized SETI@Home application I show images of where you need to look for the information in CPU-Z.

If there is nothing there, then you are not going to get a huge gain from optimization ... but, there should still be one available for that CPU.

---

CAN YOU TELL ME WHAT MB AND CPU YOU HAVE?

Who? Randy has a P3-933, no idea the MB (or any idea why it would matter...) - I have a AMD Athlon 64 3700+ "San Diego", overclocked to 2.4GHz on an ASUS A8N-E. (Plus the Macs that I really care about.) My SETI times on the AMD/Windows box, using YAOSCW-W-r8, are about one hour.

As for Randy,overclock the P3-933 to 966,then plase 5.2.2 with boing 4.72sse optimise client and then YAOSCW-Kr8.1.Now run a bench. PS: do not fall down.
As for the AMD friend ,you are in the top of things,for now!!!!

---

Thanks everyone for your help!

I'm just a happy "little" cruncher....never been worried about my rank or anything. It's a very cool project and so I do my part.

I decided to upgrade to the latest 5.22 and figured why not optimize at the same time. I've never optimized before and thought I'd try it while upgrading.

I had some troubles along the way...(you can see by this thread) and almost gave up....but I had some free time and lots of great help from you guys.

Well, the results are in. I went from 9 - 10 hour per WU crunching to 6 hours on the 1st WU and 5:45 on the next......it's unreal to me.

5.22 and the optimizer seem to be running perfectly right now.

Thanks to everyone who helped!

I'm going for the #1 spot now.....lol

"Randy"
~~~

---

[img][/img]

For Bill. :)

"EB" is meaningless.

Not really meaningless: The "B" stands for the fact that this CPU runs on a 133MHz System Bus. The "E" stands for the fact that the CPU is produced using the 0.18 micron manufacturing process.

The Pentium III has both MMX and Internet Streaming SIMD Extensions.

---

For Bill. :)

"EB" is meaningless.

Not really meaningless:

Okay, I wasn't precise in my answer... "EB" is meaningless as far as deciding which optimized client to install, unless there is some table somewhere that says "EA" doesn't have SSE, "EB" has SSE, "XY" has SSE2... but running CPU-Z will tell you all of the above, making such a table unnecessary. How's that? :-)

(AFA P2 vs P3 vs P4 etc; I'm a Mac guy. I can tell you about G3s and G4s, but not Panything!)