ASRock A780GXE/128M AM2+/AM2 AMD 780G ATX AMD Motherboard -
AMD Phenom 9600 Agena 2.3GHz Socket AM2+ 95W
OCZ2RPR10664GK (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 1066 (PC2 8500)
Western Digital Caviar SE WD3200AAJS 320GB 7200 RPM SATA 3.0Gb/s Hard Drive
# OCZ ModXStream Pro OCZ600MXSP 600W ATX12V V2.2 / EPS12V
Ok, I couldn't get one of the dimms to run at 1066 so I read on some other threads to run at 800 and overclock instead. So I have the ram running at 4-4-4-15-20 at 1.9v, CPU is at 2.5ghz and NB is at 2ghz (bumped up from 1.8). It passes memtest86+ and prime95 for half hour or more on each of the torture tests (I'll test longer later).
On Sisoft Sandra memory benchmarks, the bandwidth is about 95% of when I ran it at 1066. The latency test seems really low, the only reference systems it has to compare seem really old (Sandra version is latest), even so its running down with VIA chipsets with ddr333 and 266. Is there some compatibility problem with the latency test, should I ignore it or does this indicate some other performance problem? I'm also annoyed that I don't get as high a bandwidth as at 1066, do the tighter timings mean I'm getting some other kind of benefit not showing in the benchmark? (that would make me feel better).
I know (now) that lots of people have had trouble getting ram stable at 1066 with phenoms. I'm no engineer, but I thought the point of specifications was that if your ram, motherboard, and processor all said they supported 1066, you could run 1066. Register me as officially disgruntled.
AMD Phenom 9600 Agena 2.3GHz Socket AM2+ 95W
OCZ2RPR10664GK (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 1066 (PC2 8500)
Western Digital Caviar SE WD3200AAJS 320GB 7200 RPM SATA 3.0Gb/s Hard Drive
# OCZ ModXStream Pro OCZ600MXSP 600W ATX12V V2.2 / EPS12V
Ok, I couldn't get one of the dimms to run at 1066 so I read on some other threads to run at 800 and overclock instead. So I have the ram running at 4-4-4-15-20 at 1.9v, CPU is at 2.5ghz and NB is at 2ghz (bumped up from 1.8). It passes memtest86+ and prime95 for half hour or more on each of the torture tests (I'll test longer later).
On Sisoft Sandra memory benchmarks, the bandwidth is about 95% of when I ran it at 1066. The latency test seems really low, the only reference systems it has to compare seem really old (Sandra version is latest), even so its running down with VIA chipsets with ddr333 and 266. Is there some compatibility problem with the latency test, should I ignore it or does this indicate some other performance problem? I'm also annoyed that I don't get as high a bandwidth as at 1066, do the tighter timings mean I'm getting some other kind of benefit not showing in the benchmark? (that would make me feel better).
I know (now) that lots of people have had trouble getting ram stable at 1066 with phenoms. I'm no engineer, but I thought the point of specifications was that if your ram, motherboard, and processor all said they supported 1066, you could run 1066. Register me as officially disgruntled.