argh I vote for an "unban" in light of this constructive/informative thread
Printable View
argh I vote for an "unban" in light of this constructive/informative thread
For some reason I seem to have hit a brick wall at 262 FSB as well. I lowered every multiplier I know of, but I can't seem to get anything stable at 262 FSB, even at a cpu multiplier of 5 and my vcore maxed to 1.5625v.
I would guess i could probably get 260x11 stable @ 1.5625, which would result in 2860 mhz, and probably still managable cpu temps. I'm not going to test that right now. I don't see how I could get any higher, as the MB tops out at 1.5625V and I can't hit 262 FSB anyway.
I will be back at about 5 PM tomorrow (Mountain Time).
If you have any benchmark requests, please list them.
you won't need 1.7v just better cooling!
I can achieve a stable 2.67Ghz @ 1.35V , these board/NB clock/ whatever issues are what's preventing 3Ghz + on the right cooling
might be a hole, tried jumping 5-10mhz up from there?
Good TicTac! Should be very intresting to run Phenom on 790FX and set up the NB speed to different levels and see how the processors efficiency increases.
This was originally one of AMD's weapons I think but to run the IMC/L3 cache faster than the processor itself (async like the ALU's in the Netburst architecture).
It's surely one of the bottlenecks in the processors performance. The L3 add's about ~20ns memorylatency. Phenom looks good, but is too early and not finished, just released in a crappy form. :down:
So what? Because it's older you shouldn't flash a bios? It's provided by the manufacturer for you to flash so you could flash back and forth to no end if you are willing. I never killed a bios flashing a board either even in the NF2 days when mobos liked to die from a bios flash every now and then but that doesn't make me feel smarter as i know it's just something that can happen regardless of the user's skills and knowledge.
Live in France but that should not even be taken in consideration, as i said official bioses are meant to be flashed obviously as they are issued to fix bugs, support new cpu revs, enable functions and so on, so someone denying you warranty based on a bios flash is just screwing you IMO.
What, banned? Com'on, with this fruitfull trhead going on because of him, I'm sorry but this one of the rare cases where the reasons should be shown!
Com'on, we just lost at once 3 different motherboard full tested in phenom for what?
I agree the reason in this case is very important, can any of the previosu posts be considred reliable or not..... was he banned for faslifying posts.... or was it an old banned member....???
Intersted people need to know if any of the previous information is valid or not...???
Do we speculate and just assume that everything that AMD_Phenom posted should be ignored....????
C'mon mods, just a brief reason for the ban so people can make an educated decision on how to consider the information from earlier on in the thread....
BTW in case anyone was wondering I still havn't heard back from ASUS, still waiting for any reply.... nothing new been posted on their website either...
Yep,i just saw that Phenom dude got banned?Why was that?
We have just a few guys over here with Phenoms and now we have one less...Bad for all of us :shakes:
IF he was that dude,it doesn't mean that he never had Phenom CPU...Actually i am looking at some of his screen shots and they seem very legit.
Note that i am speaking about present posts by that dude(if it is the same dude in the first place).I don't know what he claimed before.
I heard rumors that there was going to be a black "unlocked" version of the Phenom comming out. If that's true maybe we can get some of the results that some have gotten with the ES samples. Like this one for example.
http://www.xtremesystems.org/forums/...d.php?t=166031
Edit: Dont know why the person was banned, but it is a shame we have one less Phenom to look at.
The rumor said 9600 BE in time for christmas, irrc
Could be a good point with unlocked multi for oc'ing on "old" AM2 boards, thinking of ht link speeds..
His phenom is definitely legit. My scores in Nuclearus were almost the same as his across the board.
I ran some Everest tests, but Everest seems to be highly inconsistent now, and many tests I ran yielded unrealistic results.
At some point I got 11000+ MB of memory throughput, and the next test at the exact same setting yielded like 6500.
I don't even think I can reach a solid concensus on what the differences in scores between the different NB speeds were. If there were any differences, they would be within about 100 MB/second, which far smaller than the margin of error i achieved.
CPU was clocked at 2.64 Ghz, memory at DDR2-800 at 5-5-5-12
I will say this.
About half the time I got results between the ones I'm going to list
Read - 6500-6600 MB/sec
Write - 2900-3000 MB/sec
Latency - 18-22 ns
-----------
Somehow the latency seems really low, but I constistently got it. Everest is not very well suited for this test i guess. I will have to use another program.
I ran Sandra's Memorybandwidth program, at NB multipier of 9,6, and 4, and got the results of around 6200 mb/s Int and Float SSE2 of Memory Bandwidth for each one.
I have a funny idea this northbridge multiplier thing isn't working. AMD overdrive confirms its been changed, but I never restarted, so it may have just been reading a value that was never enacted. Whenever I did restart though, the NB multiplier was back to 9 though, so I don't really understand.
Its possible that the reason I've been hitting a wall a 262 FSB was because the NB was still clocked too high, but I thought it was clocked down.
If that is the case, I never got the NB multiplier thing to work...
Let him back in.
Kick the intel trolls out.
Where the hell are all our freedoms that kids are dying for?
NB frequency does not change for me with the GIgabyte board.
The register changes, and CPU-Z picks up the new value, but you can make it anything you like.. 6ghz even. Reboot and it's gone :(
Any ideas tictac, if you're watching?
These threads are getting messy BTW
He was banned for having multiple accounts... I think last count was 5.
There is another way to show the influence of NB Speed. If you can set different CPU Multi try something like this:
CPU Multi 11 * HTT 200 = 2200 Mhz / 1800 MHz NB Speed
Memory & Cache Bench
CPU Multi 10 * HTT 220 = 2200 MHz / 2000 MHz NB Speed
Memory & Cache Bench
CPU Multi 9 * HTT 244 = 2200 MHz / 2200 Mhz NB Speed
Memory & Cache Bench
Good luck .
Ok well for anyone interested I finally had a reply of sorts from ASUS...essentially it seems that the official position is that the M2N32SLI does support Phenom from BIOS 1503, however it seems from other posts in the ASUS forum, that the BIOS locks the HT Multi at x1 regardless of what it displays in the BIOS settings screen.... this is the chipset issue thats causing the gfx performance issues, but there are also other problems with the BIOS...so for anyone with the M2N32SLI, I'd recommend against trying to use this BIOS.....
Anyway heres their reply:
http://i91.photobucket.com/albums/k2.../ASUSreply.jpg