Yep, it works exactly the same.
This MoBo is monster. I have managed to solve coolboot problems and I'm so happy, this is the best MoBo I've seen taking price into account (much better than Neo 2 by far).
Yep, it works exactly the same.
This MoBo is monster. I have managed to solve coolboot problems and I'm so happy, this is the best MoBo I've seen taking price into account (much better than Neo 2 by far).
I read a post by Bluetooth on the OCW forums that you only need to change the CPU-NB link speed, not the NB-SB link one. I think I've tested that by leaving NB-SB at default and only lowering CPU-NB, but I'll have to reboot and look in the BIOS for sure.
Under which heatsink is the northbridge? The silver or the orange one with Asrock on it? Which one is the hottest? I would assume the northbridge.
Thanks.
Thanks.
hmm so stay with the 1.2 or upgrade to 1.5? I want to be careful i killed my last board with the windows 1.4 update eventhough the bios utility showed everything was just fine. Had to RMA the board. Probably from now on i will use the dos version to update my bios.
The only immediate difference I saw between 1.2 and 1.5 was the option to set 1T command rate for RAM. Didn't seem to have any differences in overclocking ability, for the short time I used 1.5 anyway.
I never used 1.2, but with both 1.3 and 1.4 I was never able to successfully change the vcore setting. I could change it in BIOS, but it would always show as 1.4v in Windows and when I went back into BIOS, it would be reset to the default. With 1.5 I don't have that problem -- I can set vcore to 1.425 or 1.45 and it sticks.
Also, I was unable to set the CPU multiplier successfully -- same thing would happen as with vcore. With 1.5 I was finally able to set CPU multiplier to something other than 11, which has come in handy since my RAM turned out to not be stable at 436MHz and 1T (infrequent spontaneous reboots). Using CPU multiplier of 10 I can OC to 2.4GHz and keep the RAM at 400MHz/1T (using RAM divider). Oddly though, I have to set CPU-NB to 600MHz or it won't boot. 800MHz should work in theory, since with HTT at 240 it should put CPU-NB at 960, which is well within spec. But at 600 (720) it boots fine. So there may still be some problems with the BIOS, but it's definitely an improvement over 1.3 and 1.4.
I have run Prime95 for 10 hours with the above settings with 0 warnings 0 errors. Tonight I'm going to run Memtest86 overnight to make sure that 1T setting is completely stable at 400MHz.
My board came with 1.20 BIOS out of the box and I still use 1.20. There is the option to set 1T or 2T command rate in the BIOS along with other memory settings. I keep reading about this limitation but there isn't one. 1T is available in 1.20.
I'm thinking of flashing to OWC3 to see if it clears the freaky stability problems I get at 1T. I can play BF2 for hours at 1T but get unpredictable and rare reboots in memtest86+ and oddly enough, a predictably consistent reboot when attempting to paste a screenshot into paint after running benchmarks from Windows.
Well this is bizarre ... even with everything else at stock, I cannot pass Memtest86 with Command Rate set to 1T. It always fails in the same place, right after the 14 minute mark. I had thought with 2 sticks of RAM, 1T should always work at stock speeds? I also tried running the memory at CAS-3 instead of 2.5, as well as setting RAM voltage to "High" -- still failed just after 14 mins.
I can run Prime95 for 10 hours at 1T, but Memtest86 fails at 14 minutes. Very strange. I have had one random reboot at 1T though, so I am inclined to believe Memtest86.
Memory is Corsair ValueRAM 2x512 at 2.5-3-3-8 (SPD). I'm using BIOS version 1.5.
Any possible solutions, or am I just stuck at 2T on this board? I sure hope not, because 2T drops my memory bandwidth by 20% in Sandra.
I can run Prime95 for 10 hours at 1T, but Memtest86 fails at 14 minutes. Very strange. I have had one random reboot at 1T though, so I am inclined to believe Memtest86.
Memory is Corsair ValueRAM 2x512 at 2.5-3-3-8 (SPD). I'm using BIOS version 1.5.
Any possible solutions, or am I just stuck at 2T on this board? I sure hope not, because 2T drops my memory bandwidth by 20% in Sandra.
Check the Corsair forums. They'll likely have a good idea what speeds you should be able to run with that memory.
WT
Diamond Member
- Sep 21, 2000
- 4,816
- 60
- 91
Is Sandra a new game that were are unaware of ??? C'mon guys, benchmarks are one thing, game and system performance is another. This sounds like the S939 guys ragging on the S754 guys for their lack of dual channel when in fact that made maybe 6% difference in those benchmarks. We don't use the PC for benchmarks, we use it for gaming or other purposes. A synthetic benchmark seems like a poor indication of what the overall package is capable of. I hope to not sound negative about this, but its not as important as having a stable rig that doesn't crash for a simple copy and paste from an above thread.
Well obviously if I have to choose between stability and memory bandwidth, I will choose the former. I posted a request for information, not an invitation for a lecture. While it is true that synthetic benchmarks are not always indicative of system performance, 20% is quite a large percentage when talking about memory bandwidth (or anything else for that matter).
I don't see how trying to run memory at 1T at stock speeds is unreasonable. It's not like I'm complaining that something won't overclock or run out of spec. It's a pretty standard setting.
I agree that this board should run 1T, but it doesn't for me either. But there's truth in WT's somewhat sarcastic response. When I test in 3D benches and games, the 1T-2T difference is about 3%-5%. There's another thread here on AT with far more exhaustive testing which supports that general conclusion and the associated thread comparing performance at different timings is also interesting. I would say that unless something changes, you're stuck at 2T with this board and if running in that configuration has an unacceptable impact on the apps/benches that are relevant to you, it's probably not the best board for your rig. By the way, is your "14 min. mark" at the dreaded test #8? That's where mine dies at 1T.
WT
Diamond Member
- Sep 21, 2000
- 4,816
- 60
- 91
Som, it really wasn't a lecture ... thats why I added the last line so you wouldn't read it as a flame. I guess I'm just tired of people using benchmarks to indicate their rig is not as expected. I agree with you entirely on the 1T settings, but this is a known issue for this board and we all hope that it does indeed become resolved by a later BIOS update, altho Asrock will be more inclined to put more effort into the newer Uli 1697 at this point. Have you tried it with only 1 stick of RAM ?
PS - Got my Asrock board today along with my 144 .. its on the porch awaiting my arrival after work
PS - Got my Asrock board today along with my 144 .. its on the porch awaiting my arrival after work
Well if it's a known issue then I probably won't mess with it much more -- although I am tempted to try Memtest with one stick of RAM. I may do that after work. Otherwise I guess I'll just stick to 2T and try pushing the memory bus speed as high as it will go -- there's more than one way to skin a cat.
Somehow I missed that Command Rate was an issue with 2 sticks -- I couldn't run 1T when I had 3 sticks but for some reason I was under the impression that it should work with 2. Going from single-channel 2T to dual-channel 1T was a pretty enormous bandwidth increase, so it's disappointing to have to scale it back down. Still an improvement though.
Somehow I missed that Command Rate was an issue with 2 sticks -- I couldn't run 1T when I had 3 sticks but for some reason I was under the impression that it should work with 2. Going from single-channel 2T to dual-channel 1T was a pretty enormous bandwidth increase, so it's disappointing to have to scale it back down. Still an improvement though.
SynthDude: You're running an X2 right? With a dual-core proc there's no option in the BIOS 1.2 to set the 1T/2T rate. Not sure if 1.5 changes that, not that I really care if they still haven't made it stable. Question, are you back to OCWB1?
Somniferum: The 1T stability issue is specific to this board. The A64 memory controller can run 1T with 2 double-sided DIMMs. Teh ASRock however displays instability running 1T at any speed over 133, if I understand right. (Or is it at any divider other than 133? That would be good.....)
WT: In my experience I agree with the overall importance of running at 2T - it does provide a measurable performance decrease in games but not nearly to the extent that people make it out to be.
Somniferum: The 1T stability issue is specific to this board. The A64 memory controller can run 1T with 2 double-sided DIMMs. Teh ASRock however displays instability running 1T at any speed over 133, if I understand right. (Or is it at any divider other than 133? That would be good.....)
WT: In my experience I agree with the overall importance of running at 2T - it does provide a measurable performance decrease in games but not nearly to the extent that people make it out to be.
I do just got out of my AsRock, but i do not see why everyone thinks there is a COMPLETE issue with 1T. But i never had one issue with this ram and being in 4 different mobos (i875,nf3, M1695, and nf4) always worked with stock timing 1T with 10-30Mhz increase in spd. I guess go Corsair lol.
I never ran into any issues running 1T with anybios 1.2-1.5 and OCWBeta bios. This was with 270x9 with a 166 Memory Divider. The parts are all the same listed in my sig other than the DFI board doh.
Just in case some one wants to know why i have a DFI now, well because i found the bios NEVER was stable for anything above 274 no matter what voltage the core had. There was some little quirks i did not like but overall if you are not going to go past 274Mhz FSB then this board is great for the money and Fast. I got the same exact PCMark05 score with each board and parts listed in sig +-1%.
So now that i do not own the board anymore (giving it to my sister with a Venice 3000+), i will like to resay the board is fast and good for any speeds from 200-274 FSB.
I never ran into any issues running 1T with anybios 1.2-1.5 and OCWBeta bios. This was with 270x9 with a 166 Memory Divider. The parts are all the same listed in my sig other than the DFI board doh.
Just in case some one wants to know why i have a DFI now, well because i found the bios NEVER was stable for anything above 274 no matter what voltage the core had. There was some little quirks i did not like but overall if you are not going to go past 274Mhz FSB then this board is great for the money and Fast. I got the same exact PCMark05 score with each board and parts listed in sig +-1%.
So now that i do not own the board anymore (giving it to my sister with a Venice 3000+), i will like to resay the board is fast and good for any speeds from 200-274 FSB.
Banshee couple questions about your RAM. 1) did you run memtest86 test #8? That and pasting a JPG into paint usually are telltale signs of instability at 1T.
2)did you try the OCWB1 bios when pushing your overclock? I've heard that 1.2+ won't work over 270FSB.... some people have had teh board up to 400. No I don't have the link right now
2)did you try the OCWB1 bios when pushing your overclock? I've heard that 1.2+ won't work over 270FSB.... some people have had teh board up to 400. No I don't have the link right now
I used the system since the board came out what a couple of months did alot of paste copying stress testing i am positive i had a stable system with 1T at 270FSB.
Yes i tired OCB1 still unstable.
I was able to push the FSB to what ever i wanted to in windows to 2.75Ghz with a64tweaker and a64info, but these programs have their own bugs and that why i said Bios Stable.
I do not know if something with my hardware did not like to go over 275FSB stable but it wasn't.
For example i was able to boot with 300x9 ONLY at 100Mhz divider, anything higher than that woudl result in a HardDisk error that looked like it wasn't locked anymore. again have no clue but when i can only use the 100Mhz divider to me is what i call unstable. This is with 1.55-1.6vCore.
Did not really try much above 300 as using lower multiplier seemed to be unstable on these bios too (only tried 1.2) but changing the multiper with high FSB 330+ would result in what many call coldboot problems where when restarted it would not post, but sometimes it does. Again i call unstable bios
I am not quite sure how many people use over 274 FSB daily on this board not just a screen shot, so i am not sure if most of those people can say it was complete stable. My issues with the board was not as much as windows was unstable just posting was like a flip of a coin when bios is over 274.
Good Luck everyone.
Yes i tired OCB1 still unstable.
I was able to push the FSB to what ever i wanted to in windows to 2.75Ghz with a64tweaker and a64info, but these programs have their own bugs and that why i said Bios Stable.
I do not know if something with my hardware did not like to go over 275FSB stable but it wasn't.
For example i was able to boot with 300x9 ONLY at 100Mhz divider, anything higher than that woudl result in a HardDisk error that looked like it wasn't locked anymore. again have no clue but when i can only use the 100Mhz divider to me is what i call unstable. This is with 1.55-1.6vCore.
Did not really try much above 300 as using lower multiplier seemed to be unstable on these bios too (only tried 1.2) but changing the multiper with high FSB 330+ would result in what many call coldboot problems where when restarted it would not post, but sometimes it does. Again i call unstable bios
I am not quite sure how many people use over 274 FSB daily on this board not just a screen shot, so i am not sure if most of those people can say it was complete stable. My issues with the board was not as much as windows was unstable just posting was like a flip of a coin when bios is over 274.
Good Luck everyone.
Yep, the dreaded test #8 it is. Damn you, Modulo 20 Ones & Zeroes!!!
I tried running Memtest86 in some different RAM configurations and found that either DIMM, by itself, passes all tests at 1T with flying colors. So it would seem to be an issue with 1T + Dual Channel mode. I tried proving this once and for all by using DIMM1 and DIMM3, running both sticks in Single Channel, but for some reason the board won't POST in this configuration (although it should work according to the ASRock manual).
Of course at that point I had to wonder: If I could get it to POST in Single Channel 1T and it was stable, would it be worth it? After running some tests, at least from a pure memory bandwidth perspective, the answer would appear to be NO:
200MHz Single Channel, 1T: 3031MB/s
200MHz Single Channel, 2T: 2716MB/s
200MHz Dual Channel, 1T: 5776MB/s
200MHz Dual Channel, 2T: 4710MB/s
Dual Channel 2T is actually 55% faster (!) than Single Channel 1T, in terms of pure memory bandwidth. (Measured with Sandra -- apologies to the anti-synthetic benchmark crowd.) I'm sure that will come as no surprise to most, but to an AMD/dual channel newb like myself it was pretty interesting to see the difference.
I also tried different Dual Channel + 1T configurations -- swapping DIMM1 and DIMM2, or using DIMM3 and DIMM4 instead -- but test #8 got me every time.
So I was completely convinced that it's an issue with Dual Channel + 1T on this board, until I read Lord Banshee's post. Now I'm wondering about memory dividers. Maybe it's time for more testing....
WT
Diamond Member
- Sep 21, 2000
- 4,816
- 60
- 91
Som, thanks for those test numbers ... every little bit of info helps the Asrock owners with their own setups. If that is your conclusion, can it also be said that this is also a RAM issue, with certain sticks running 1T and others not so successful ? We may find a trend with certain sticks being better suited to the board and even the particular BIOS. I have a pair of cheap Kingmax 3500 sticks to use, but would like to get a G.Skill 3200 1gb stick, as that seems to be a good 'performance' stick (ie: good buy for the buck).
Well, both of my sticks work at 1T individually, so to me it seems more like a Dual Channel + 1T issue than a RAM issue. But Lord Banshee's results seem to contradict that. What's interesting is that Lord Banshee and I are both running the exact same RAM -- Corsair ValueRAM @ 2.5-3-3-8. I tried fooling with the memory divider, but at 166 (240*10) test #8 crashed as usual at 1T. I tried using 133 (270*9) but I couldn't even POST with that setup, even with CPU-NB all the way down to 2.
I'd be interested to know if anyone has trouble running 1T with a Single Channel setup -- like that 1Gb stick you mentioned. If my theory is right then it should work just fine at 1T.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter