2700+ Barton vs 2700+ T\'Bred

Author	Message
old_user16598 Send message Joined: 11 Sep 04 Posts: 9 Credit: 321,368 RAC: 0	Message 6484 - Posted: 30 Nov 2004, 8:47:58 UTC I have these two machines setup running CPDN and both are overclocked somewhat. Funny thing is the T'Bred seems to get higher CPDN benchmark numbers but is consistently slower by a good margin in trickle times. Is the benchmark not very reliable or what? Keith ID: 6484 · Reply Quote

Andrew Hingston Volunteer moderator Send message Joined: 17 Aug 04 Posts: 753 Credit: 9,804,700 RAC: 0	Message 6486 - Posted: 30 Nov 2004, 12:43:15 UTC Is the benchmark you mean the one calculated by BOINC? It is devised by the BOINC team and used in calculting credits for S@H and some other projects. It isn't used by CPDN, and experience is certainly that the HADSM3 model does favour some processors over others. Trickles are what count, and are the only valid way of comparing machines running BOINC. ID: 6486 · Reply Quote

old_user16598 Send message Joined: 11 Sep 04 Posts: 9 Credit: 321,368 RAC: 0	Message 6488 - Posted: 30 Nov 2004, 13:45:39 UTC Yeah, that is what I was referring to. Well, the larger cache of the Barton must really be worth alot. It is doing trickles in 2.5 vs 3.1 sec for the T'Bred. ID: 6488 · Reply Quote

old_user2855 Send message Joined: 29 Aug 04 Posts: 2 Credit: 58,355 RAC: 0	Message 6656 - Posted: 6 Dec 2004, 21:39:28 UTC The benchmark is built into the BOINC client, it is not CPDN related and not really accurate anyway. Btw my barton trickles around every 3.2 sec. Must be the sis chipset. >. ID: 6656 · Reply Quote

old_user7130 Send message Joined: 31 Aug 04 Posts: 3 Credit: 546,392 RAC: 0	Message 6666 - Posted: 7 Dec 2004, 2:10:42 UTC - in response to Message 6488. > Yeah, that is what I was referring to. Well, the larger cache of the Barton > must really be worth alot. It is doing trickles in 2.5 vs 3.1 sec for the > T'Bred. >Yes, the benchmarks are more or less useless. It seems to me that effective memory speed is critical. Does your Barton have a 333 FSB and the Thoroughbread a 256? ID: 6666 · Reply Quote

old_user16598 Send message Joined: 11 Sep 04 Posts: 9 Credit: 321,368 RAC: 0	Message 6818 - Posted: 9 Dec 2004, 20:22:51 UTC - in response to Message 6666. They are both 333MHz FSB ID: 6818 · Reply Quote

old_user7130 Send message Joined: 31 Aug 04 Posts: 3 Credit: 546,392 RAC: 0	Message 6828 - Posted: 9 Dec 2004, 23:05:21 UTC - in response to Message 6818. > They are both 333MHz FSB > Here are some numbers for you. In our family team we have two PCs, both Windows XP home, both AMD Athlon 2000+ (Palomino 256FSB) (not overclocked) and both receiving similar benchmarks. But one is knocking trickles out in 3.4 s/TS and the other at only 4.7. But there is one major difference between the two. The slower machine has 256 memory because that is the fastest the motherboard can be set to and the second has 333 memory. We also have an AMD64 3000+ PC being overclocked about 16% also with 333 memory. It is doing about 2.06 s/TS. If you interpolate between the two 333 machines I think you will find that your Barton is doing about right. If your T'bred has a 333 FSB and 333 memory then your question should really be why it's not producing faster trickles. ID: 6828 · Reply Quote