Here is attempted to collect overclocking results for 940, 754, 939 CPU's and systems.
Please post A64 overclocking results only, and refrain from discussions and questions to individual.
Result tabulations for
754 ClawHammer, NewCastle
939 FX, Winchester, NewCastle
939 Venice, San Diego, X2 Toledo, X2 Manchester, Opterons
Dual core Manchester/Toledo overclocking estimation from Venice/SanDiego statistic
Different core performance results in SuperPI 32M (post #3 and #4)
Analysis of different core performance in SuperPI 32M (post #308)
Please complete according to the following format for consistency, so results can be used for general comparison.
This cut and paste link brings a cut and paste format for posting your result.
You may also add your benchmarks, techniques and addons (such as vmod, heatsinks to MOSFET, chipset, ...), comments, links to your oc threads, ....
To make your results and this thread more creditable, it is recommended to attach screen shots such as
- cpu-z or gcpuid or wcpuid or alike for bootable/stable frequency
- stable frequency: prime95, SuperPI 32M, 3DMark01, (3DMark03) for stable frequency
(The stable frequency posted should be at least passing SuperPI 32 M, 3DMark01 and 3DMark03, preferably Prime95 12-24 hours)
- link to your existing screen shots (somewhere else) with brief description how stability was tested.
It is better to post multiple entries (one per post) in case you
- have hardware change (e.g. cooling, motherboard, memory) for comparison
- have multiple setups
When you have new result(s) with same hardware, you may append or edit the original post.
In case your result has changed, not been tabulated, or was tabulated incorrectly, PM me.
model number: ADA3000AEP4AX <- (if you don't know, look at this A64 CPU link)
core description: NC/CH/SH C0/CG desktop/mobile ... <- (optional, for checking, clarity)
5-letter code: CBASC
week-code: 0416
4-letter code: XPMW
13-letter code: 9833965C40135 <- (optional if you don't have it)
CPU frequency: highest stable 2610 MHz (261 x 10), optional highest boot <- (don't forget HTT and CPU multiplier)
CPU voltage: 1.65 V
temperature (system idle/load, CPU idle/load): system 26/28 C, CPU 40/48 C
cooling: SLK-948U TT SFII <- (air or water or others, with some brief details)
motherboard: XYZ rev 1.1 bios 1.23 <- (revision, bios if relevant)
memory: XYZ PC3700 512 MB x 2
memory bus frequency: 261 MHz
memory timing: 2.5-4-4-8 1T <- (tCAS-tRCD-tRP-tRAS, don't forget 1T/2T)
memory voltage: 2.8 V
HT frequency: HTT x 4 <- (Usually between 3x - 5x)
power supply: XYZ 500 W
From time to time, I may summarize and tabulate cumulative results, and hopfully some trends may emerge.
Your result may not be tabulated if there is missing or inconsistent information.
Need more data points for Venice and SanDiego.
With more data, more analysis can be run to show what is going on than guessing, and in turn helps us all.
E.g. with the data, this is what analysis can do.
- It shows 90 nm CPU can be clocked higher with lower voltage.
- Venice/SanDiego can be clocked higher, about 200 MHz at least on average than Winchester, NewCastle, ClawHammer.
- It shows the 90 nm SOI processors (e.g. SanDiego/Venice/Winchester) can be clocked at higher frequencies while running at lower voltages, compared to that of the 130 nm (NewCastle/ClawHammer).
- The 90 nm SOI with DSL strained silicon (e.g. SanDiego/Venice) further improves the frequencies (by 200+ MHz on the average) compared to 90 nm SOI (non DSL) (e.g. Winchester).
- With more data, we can do analysis on steppings such as CBBLE vs LBBLE, CBBFD/CBBHD vs CBBID, CABHE vs CABGE.
Already have some interesting data, but need to hold off publishing them until there are more data points.
e.g.
Credit:
Thanks to all taking time to enter and update results, hope this would help us all on overclocking expectation, ....
Though I have been collecting these data before creating the thread,
Buhammot prompted me via a PM to create this thread, and
dreIU suggested to have it stickied.
Jhatfie suggested posting stability screen shots.
Super Nade suggested to add CPU temperature reading.
Appendix
Revisions and steppings (under construction) (post 6)
940, 754, 939 CPU models and specifications (post 5)
How to identify the physical core of an A64
How to tell memory is running 1T or 2T (preliminary)
In case the bios does not show the choice of 1T or 2T explicitly, you may use Sandra memory bandwidth to determine 1T or 2T by looking at the memory bandwidth efficiency in the memory bandwidth benchmark result as follow:
For 754, if the memory efficiency is above 90%, consider it as running 1T;
if the memory efficiency is in the low 80%, consider it as running 2T.
For 939, if the memory efficiency is above 85%, consider it as running 1T.
Please post A64 overclocking results only, and refrain from discussions and questions to individual.
Result tabulations for
754 ClawHammer, NewCastle
939 FX, Winchester, NewCastle
939 Venice, San Diego, X2 Toledo, X2 Manchester, Opterons
Dual core Manchester/Toledo overclocking estimation from Venice/SanDiego statistic
Different core performance results in SuperPI 32M (post #3 and #4)
Analysis of different core performance in SuperPI 32M (post #308)
Please complete according to the following format for consistency, so results can be used for general comparison.
This cut and paste link brings a cut and paste format for posting your result.
You may also add your benchmarks, techniques and addons (such as vmod, heatsinks to MOSFET, chipset, ...), comments, links to your oc threads, ....
To make your results and this thread more creditable, it is recommended to attach screen shots such as
- cpu-z or gcpuid or wcpuid or alike for bootable/stable frequency
- stable frequency: prime95, SuperPI 32M, 3DMark01, (3DMark03) for stable frequency
(The stable frequency posted should be at least passing SuperPI 32 M, 3DMark01 and 3DMark03, preferably Prime95 12-24 hours)
- link to your existing screen shots (somewhere else) with brief description how stability was tested.
It is better to post multiple entries (one per post) in case you
- have hardware change (e.g. cooling, motherboard, memory) for comparison
- have multiple setups
When you have new result(s) with same hardware, you may append or edit the original post.
In case your result has changed, not been tabulated, or was tabulated incorrectly, PM me.
model number: ADA3000AEP4AX <- (if you don't know, look at this A64 CPU link)
core description: NC/CH/SH C0/CG desktop/mobile ... <- (optional, for checking, clarity)
5-letter code: CBASC
week-code: 0416
4-letter code: XPMW
13-letter code: 9833965C40135 <- (optional if you don't have it)
CPU frequency: highest stable 2610 MHz (261 x 10), optional highest boot <- (don't forget HTT and CPU multiplier)
CPU voltage: 1.65 V
temperature (system idle/load, CPU idle/load): system 26/28 C, CPU 40/48 C
cooling: SLK-948U TT SFII <- (air or water or others, with some brief details)
motherboard: XYZ rev 1.1 bios 1.23 <- (revision, bios if relevant)
memory: XYZ PC3700 512 MB x 2
memory bus frequency: 261 MHz
memory timing: 2.5-4-4-8 1T <- (tCAS-tRCD-tRP-tRAS, don't forget 1T/2T)
memory voltage: 2.8 V
HT frequency: HTT x 4 <- (Usually between 3x - 5x)
power supply: XYZ 500 W
From time to time, I may summarize and tabulate cumulative results, and hopfully some trends may emerge.
Your result may not be tabulated if there is missing or inconsistent information.
Need more data points for Venice and SanDiego.
With more data, more analysis can be run to show what is going on than guessing, and in turn helps us all.
E.g. with the data, this is what analysis can do.
- It shows 90 nm CPU can be clocked higher with lower voltage.
- Venice/SanDiego can be clocked higher, about 200 MHz at least on average than Winchester, NewCastle, ClawHammer.
- It shows the 90 nm SOI processors (e.g. SanDiego/Venice/Winchester) can be clocked at higher frequencies while running at lower voltages, compared to that of the 130 nm (NewCastle/ClawHammer).
- The 90 nm SOI with DSL strained silicon (e.g. SanDiego/Venice) further improves the frequencies (by 200+ MHz on the average) compared to 90 nm SOI (non DSL) (e.g. Winchester).
- With more data, we can do analysis on steppings such as CBBLE vs LBBLE, CBBFD/CBBHD vs CBBID, CABHE vs CABGE.
Already have some interesting data, but need to hold off publishing them until there are more data points.
e.g.
Credit:
Thanks to all taking time to enter and update results, hope this would help us all on overclocking expectation, ....
Though I have been collecting these data before creating the thread,
Buhammot prompted me via a PM to create this thread, and
dreIU suggested to have it stickied.
Jhatfie suggested posting stability screen shots.
Super Nade suggested to add CPU temperature reading.
Appendix
Revisions and steppings (under construction) (post 6)
940, 754, 939 CPU models and specifications (post 5)
How to identify the physical core of an A64
How to tell memory is running 1T or 2T (preliminary)
In case the bios does not show the choice of 1T or 2T explicitly, you may use Sandra memory bandwidth to determine 1T or 2T by looking at the memory bandwidth efficiency in the memory bandwidth benchmark result as follow:
For 754, if the memory efficiency is above 90%, consider it as running 1T;
if the memory efficiency is in the low 80%, consider it as running 2T.
For 939, if the memory efficiency is above 85%, consider it as running 1T.
Last edited: