Oops, I have a funny story to tell.
I was actually a little disappointed in the 2:23:26 time I posted for the Ars Technica Benchmark WU run on my Pentium 4 machine. I had seen WUs with 0.417 ARs being posted better than that and upon closer inspection I found that, sure enough, I did.
The strange part of this is that my Gigabyte system with the Pentium 4 turns in the fastest "sweet" WUs by far with a string of 1:54 CPU times but usually it can only do WUs with 0.417 in around 2.21 - 2:23. When I checked the HISTORY of the TeAm Smokeball Q I found quite a few 0.417 AR WUs done in 2:19 but none by my super-duper Pentium 4 system. The fast 0.417 AR WUs were being done by my Asus System with the 2400+ XP! :Q
So I decided to run the Ars Technica Benchmark on that computer and when it was done with a time just over 2:19 I submitted the results. Even though it would probably show up as a Duplicate WU I TRANSMITTED it to Berkelely through the TS Q. Then when I looked at the Q's history I was shocked ... there was no DUP and there was no 0.417 AR WU. The WU the 2400+ XP system submitted had an AR of 0.618 and a CPU time of 2:20?
2003 Jan 11 09:00pm 6.14,11.31 0.618 Smoke #9 AsusA7N8X 2400XP 2:20
I had crunched the wrong WU ! Then I double-checked the Pentium 4 submission and, even though it was a 0.417 AR WU, upon closer inspection I discovered it too was not the Ars Technica Benchmark. Now I was completely embarrassed.
I had used a new install of SetiDriver to implement the running of these Benchmarks. I downloaded the Benchmark and unzipped it into the folder where only the SetiDriver.exe and the Setiathome3.03 files resided. I put a "1" in the cache of SetiDriver, set it to not transmit automatically and then hit TRANSMIT. I assumed SetiDriver would
not download a new WU but would just start crunching the Benchmark WU I had already placed in the folder. WRONG! It downloaded a new one from the Q overwriting the Benchmark WU and started crunching.
So now I had to write Roelof@ArsTechnica and apologize for turning in Benchmark scores that were not the actual Benchmark WUs.
But before I ?fessed up to you here on the AT Forums, I decided to run the
TRUE Ars Technica Benchmark Wu to see what my computers could actually do.
The results were even better.
And I double-checked and made sure, this time, I had actually run the correct
Ars Benchmark WU.
And here it is: 2003 Jan 12 02:32am 8.074,18.25 0.417 Smoke #9 AsusA7N8X 2400XP 2:19
The actual time including seconds was
2:19:19
Now I would like someone to explain to me the current odd situation. I have a Pentium 4 that does "sweet" WUs in 1:54 but can only do regular 0.417 AR WUs in 2:23
YET I have a 2400+ XP system that does "sweet" WUs in 2:01 but kills regular 0.417 WUs in 2:19 ? :Q