Hi you guys!
After long heavy tests I finally got it, my 3200+ Winnie caps at 2.6 Ghz, that?s 4100+ according to AMD. Everest Home Edition also recognises it as 4100+!
My current BIOS settings are these:
HTT: 289 Mhz
CPU Multiplier: 9x
CPU Vcore: 1.450v
LDT Multiplier: 3x
Ram Max Memclock (Divider): DDR333
Ram Timings: 2-3-3-7 1T running at 236.5 Mhz (PC 3800)
Note: I do not use any ASUS/nVIDIA software for overclocking such as nTUNE or AiBooster. I have not even installed them, the best results are achieved with the Bios config.
I have been working with my Asus A8N SLI Deluxe and my Winnie for a time at 2.4 Ghz (240 HTT x 10, Vcore:1.4v) in 1:1 ratio with my Ram running also at 240 Mhz 2.5-3-3-7 1T.
In Sisoft Sandra 2005 SR1 the results obtained in the Memory Bandwidth Benchmark, with this config, were these: 6683/6609 Mb/s.
I was afraid of a lower memory performance with my new configuration because the Ram memory was running at a lower speed (236.5 instead of 240) and, what?s more, I was using a divider (DDR333 instead of 1:1 ratio or DDR400). So, after getting my system stable trough Prime95 blend test and SuperPi 16M, and what I call the SWAT4 test, (enough for testing stability issues) I decided to test its performance against the results I obtained with my older config. Results impressed me, they were better than I expected: 6974/6897 Mb/s.
Hope to post later on more results in order to offer you a detailed performance comparison but I cannot right now, gotta study hard! If you want me to run any software for testing, please tell me its name and I?ll try to post my results obtained with that program so that you can compare them to yours!
Using the DDR 333 divider has no negative impact in performance at all on these platforms! I would recommend those who are running their rigs at 2.4 Ghz 1:1 ratio to try the DDR333 divider and post here their results. I am convinced that most of you should manage to get 2.6 Ghz stable! Give it a try men!
Experiences: At first (time ago) I started to run my Winnie at 2.4 Ghz 1:1 ratio (240x10) with 1.425v because I thought that voltage increase would be necessary and my Ram memory at 240 Mhz 2-3-3-7 1T since Memtest86 v3.2 reported no errors after several Passes. The interesting fact is the system appeared to be stable but Prime95 reported errors and SuperPi 4M,8M,16M,32M reported errors too. At first I did not pay much attention to these errors but I did when I bought the SWAT4 videogame; After playing this game for about 15-20 mins, it suddenly threw me to Windows everytime, no exception! I wondered why several times, I reinstalled it and the same happened, I got another copy of the game and the same story, so I decided that had to be an overclocking problem! After research, I discovered the following: at 240 Mhz my ram memory could not run stable on the ASUS with 2-3-3-7 1T timings (it seems the A8N SLI Dlx does not like ?2? timings) so I lowered the timings to 2.5-3-3-7 1T and?it ran stable on the ASUS. Then I continued experiencing some issues when testing?I remembered that my Winnie, when I first got it, could run at 2.4 Ghz with the standard voltage of 1.4v but I did never test that so I lowered the Vcore to 1.4v and?.The system ran for the first time truly stable trough Prime95, SuperPi and SWAT4, at least!!
Now, I am able to run my memory at 2-3-3-7 timings at 236.5 Mhz with no issues. Also I have been forced to rise the Vcore to 1.450v as I am running my Winnie at 2.6 Ghz, If I low the voltage the system does not run rock-solid. So for 2.4 Ghz, 1.4v and for 2.6 Ghz, 1.450v.
Note that in Sisoft Sandra 2005 in the ?CPU & BIOS Information? module > ?Min/Max Core Voltage?, it reports Min:1.400v, Max:1.450v. This may be interesting since it reported the same when I was running my winnie at 1.400v!! The question is, Is Sisoft Sandra able to detect the maximum Cpu voltage allowed by each CPU??
I would like you to try this and post here your experiences! It could be helpful in order to know the exact voltage each Cpu allows!
Best wishes,
MADMAX23
After long heavy tests I finally got it, my 3200+ Winnie caps at 2.6 Ghz, that?s 4100+ according to AMD. Everest Home Edition also recognises it as 4100+!
My current BIOS settings are these:
HTT: 289 Mhz
CPU Multiplier: 9x
CPU Vcore: 1.450v
LDT Multiplier: 3x
Ram Max Memclock (Divider): DDR333
Ram Timings: 2-3-3-7 1T running at 236.5 Mhz (PC 3800)
Note: I do not use any ASUS/nVIDIA software for overclocking such as nTUNE or AiBooster. I have not even installed them, the best results are achieved with the Bios config.
I have been working with my Asus A8N SLI Deluxe and my Winnie for a time at 2.4 Ghz (240 HTT x 10, Vcore:1.4v) in 1:1 ratio with my Ram running also at 240 Mhz 2.5-3-3-7 1T.
In Sisoft Sandra 2005 SR1 the results obtained in the Memory Bandwidth Benchmark, with this config, were these: 6683/6609 Mb/s.
I was afraid of a lower memory performance with my new configuration because the Ram memory was running at a lower speed (236.5 instead of 240) and, what?s more, I was using a divider (DDR333 instead of 1:1 ratio or DDR400). So, after getting my system stable trough Prime95 blend test and SuperPi 16M, and what I call the SWAT4 test, (enough for testing stability issues) I decided to test its performance against the results I obtained with my older config. Results impressed me, they were better than I expected: 6974/6897 Mb/s.
Hope to post later on more results in order to offer you a detailed performance comparison but I cannot right now, gotta study hard! If you want me to run any software for testing, please tell me its name and I?ll try to post my results obtained with that program so that you can compare them to yours!
Using the DDR 333 divider has no negative impact in performance at all on these platforms! I would recommend those who are running their rigs at 2.4 Ghz 1:1 ratio to try the DDR333 divider and post here their results. I am convinced that most of you should manage to get 2.6 Ghz stable! Give it a try men!
Experiences: At first (time ago) I started to run my Winnie at 2.4 Ghz 1:1 ratio (240x10) with 1.425v because I thought that voltage increase would be necessary and my Ram memory at 240 Mhz 2-3-3-7 1T since Memtest86 v3.2 reported no errors after several Passes. The interesting fact is the system appeared to be stable but Prime95 reported errors and SuperPi 4M,8M,16M,32M reported errors too. At first I did not pay much attention to these errors but I did when I bought the SWAT4 videogame; After playing this game for about 15-20 mins, it suddenly threw me to Windows everytime, no exception! I wondered why several times, I reinstalled it and the same happened, I got another copy of the game and the same story, so I decided that had to be an overclocking problem! After research, I discovered the following: at 240 Mhz my ram memory could not run stable on the ASUS with 2-3-3-7 1T timings (it seems the A8N SLI Dlx does not like ?2? timings) so I lowered the timings to 2.5-3-3-7 1T and?it ran stable on the ASUS. Then I continued experiencing some issues when testing?I remembered that my Winnie, when I first got it, could run at 2.4 Ghz with the standard voltage of 1.4v but I did never test that so I lowered the Vcore to 1.4v and?.The system ran for the first time truly stable trough Prime95, SuperPi and SWAT4, at least!!
Now, I am able to run my memory at 2-3-3-7 timings at 236.5 Mhz with no issues. Also I have been forced to rise the Vcore to 1.450v as I am running my Winnie at 2.6 Ghz, If I low the voltage the system does not run rock-solid. So for 2.4 Ghz, 1.4v and for 2.6 Ghz, 1.450v.
Note that in Sisoft Sandra 2005 in the ?CPU & BIOS Information? module > ?Min/Max Core Voltage?, it reports Min:1.400v, Max:1.450v. This may be interesting since it reported the same when I was running my winnie at 1.400v!! The question is, Is Sisoft Sandra able to detect the maximum Cpu voltage allowed by each CPU??
I would like you to try this and post here your experiences! It could be helpful in order to know the exact voltage each Cpu allows!
Best wishes,
MADMAX23
Facebook
Twitter