- Joined
- Feb 27, 2009
- Location
- Dorado, PR
I think it would be a great idea for us to share what fsb/speed/timing settings we have our memory setups running at and what our synthetic bechmarks are reporting as far as latencies and bandwidth. We can also document our observations while overclocking our memory as to aid everyone in general in their quest to tweaking and setting up their memory configurations. This is arguably the toughest part of any good stable O/C so this could become a good repository of what configurations work in which ways. We could also share a brief summary of our experiences with our memory modules as no two memory modules behave alike and some particular chips like to be setup in particular ways.
It would be a good way for all of us to compare results and have an idea if our setups are running optimally or if there is something wrong with our configurations.
I for one have the following results, lets try and follow this template for organization's sake:
Memory Modules: Mushkin HP3-1333
Motherboard/Chipset: Asus StrikerII Extreme / nVidia 790i Ultra SLI
Module Size: 2x2Gb = 4Gb
Stock FSB/Multiplier: 1333mhz / 1:1
Rated Speed: 1333mhz
Rated Timings: 7-7-7-20-1T-33tRC
Rated Voltage: 1.7v - 1.8v
Overclocked FSB/Multiplier: 2000mhz / 5:4
Overclocked Speed: 1600mhz
Overclocked Timings: 9-9-9-28-2T-35tRC
Overclocked Voltage: 1.6v
Overclocking Observations: As the speed increases this RAM prefers to be undervolted rather than overvolted. No matter how many volts you push through it it will generate errors so there really is no way to try and lower the timings as the speed increases by overvolting. This setup yielded the same bandwidth as running 2000mhz FSB with a 3:2 divider (1333mhz for the RAM) with the stock rated timings. Oddly enough it also resulted in the same latencies. Overvolting at stock speeds also did not permit tighter timings.
For now I'll post my bechmarks as plain text because I don't have my screenshots with me at the moment, I'll post screenshots of the benchies tonight. Please post benchmarks as screenshots along with a screenshot of both the Main and Memory CPUz Tabs to certify the settings and results.
SiSoft Sandra Bandwidth: 10.1Gb/s Int | 10.05Gb/s Float
SiSoft Sandra Latency: 71ns
to all!
EDIT: To any admin, I just realized I posted this in the wrong forum. Was meant to be posted in the Memory forum, can this be moved please?
It would be a good way for all of us to compare results and have an idea if our setups are running optimally or if there is something wrong with our configurations.
I for one have the following results, lets try and follow this template for organization's sake:
Memory Modules: Mushkin HP3-1333
Motherboard/Chipset: Asus StrikerII Extreme / nVidia 790i Ultra SLI
Module Size: 2x2Gb = 4Gb
Stock FSB/Multiplier: 1333mhz / 1:1
Rated Speed: 1333mhz
Rated Timings: 7-7-7-20-1T-33tRC
Rated Voltage: 1.7v - 1.8v
Overclocked FSB/Multiplier: 2000mhz / 5:4
Overclocked Speed: 1600mhz
Overclocked Timings: 9-9-9-28-2T-35tRC
Overclocked Voltage: 1.6v
Overclocking Observations: As the speed increases this RAM prefers to be undervolted rather than overvolted. No matter how many volts you push through it it will generate errors so there really is no way to try and lower the timings as the speed increases by overvolting. This setup yielded the same bandwidth as running 2000mhz FSB with a 3:2 divider (1333mhz for the RAM) with the stock rated timings. Oddly enough it also resulted in the same latencies. Overvolting at stock speeds also did not permit tighter timings.
For now I'll post my bechmarks as plain text because I don't have my screenshots with me at the moment, I'll post screenshots of the benchies tonight. Please post benchmarks as screenshots along with a screenshot of both the Main and Memory CPUz Tabs to certify the settings and results.
SiSoft Sandra Bandwidth: 10.1Gb/s Int | 10.05Gb/s Float
SiSoft Sandra Latency: 71ns
to all!EDIT: To any admin, I just realized I posted this in the wrong forum. Was meant to be posted in the Memory forum, can this be moved please?
Last edited:
