Obercoquin?? ezo ke eh?? Me echáis una mano?

[Tassadar]

Hello, friends.

I've been tinkering with computers for years, like many of you, and quite a few of them practicing overclocking. I ask people who know about this to read the post and see if they can help. Please.

My first OC was a K6-2 350@400, and the second (more serious, with a good cooler and all xD) was an early Athlon 1000, which I took up to a terrifying 1300 mhz. It wasn't my computer (I wish it was). That said, afterwards I overclocked my Duron 750@950, and later, thanks to a vcore-mod, 1030 mhz.

Well, right now I feel like a complete idiot. I don't know if I'm too old or if the current boards are a bunch of crap because they do everything automatically and when you try to overclock them, they completely fail.

The story is that I've been trying to overclock with an Asus M5A88-M for several days, and I can't do it. Don't think I'm trying to put the CPU at 7 ghz... no, I'm just trying to put the CPU below its speed (lowering the multiplier) but raise the bus and lower the memory, so I can see when the bus arrives and the next step would be to raise the CPU.

Back in the days of Socket 939, raising the bus by 20% was child's play: you entered the BIOS, put 240 instead of 200, lowered the memory from 400 to 333 mhz, and lowered the two HT multipliers from 5 to 4 and... voila! you had the memory running at the original 400 mhz, the buses at the original 2000, but we managed to raise the CPU by 20%.

I can tell you that with this board I'm not able to do that! In the manual overclocking profile, at more than 220 bus, the system doesn't even post if I leave the memory at 1333 (so that when it reaches 240 bus it goes to 1600 which is what they are). I've copied all the latencies, all that the board allows to configure, and I've configured them manually for those that it picks up when everything is set to default. This is because if I put 1333 in the BIOS, it picks up the latencies of 1333, and it doesn't boot putting 240 mhz bus even though the memory goes to 1600. The most I've managed is, putting the latencies manually, that Windows boots and everything, but a few minutes of linpack make it error out.

So I tried using a feature of the BIOS (D.O.C.P. or something like that) that seems to fix the latencies automatically to the SPD of 1600 mhz, so that then you can manually put 1333, and raise the bus to 240 without worrying about the memory. In theory it's great, it easily allows you to increase the bus to 240 to raise the CPU but keep the memory with its original latency and speed, right?

BUT NO, it's not stable at all, I've checked with AMD overdrive and the latencies are OK (they put the SPD 1600's own), but it doesn't work, and don't think it fails because I'm raising something else, no! I've lowered the CPU and NB multipliers, so the CPU is below its original 3000 mhz, the NB is below 2000 (its original speed) and the HT bus, going in sync with the NB, is also below 2000.

In total, what I do is put everything in a way that has 240 bus, but no COMPONENT is working above its speed. As I say, it doesn't work, in the last test the linpack crashed after a few minutes (what the hell).

What am I missing? What am I doing wrong?

The "tricky" thing is that today I've sent a board to pccomponentes for RMA, not a 20 euro one, no, a Gigabyte GA-970A-UD3 AM3+... what was wrong with it? Well, nothing, as simple as that, I've tested two Phenom II x2 that are 100% stable and functional on other mutated boards to x4, and on this board, after 10 seconds of linpack, the machine hangs, even everything by default and even lowering the CPU to 2.2 ghz (for the whole face, besides I've done a billion tests and it's proven that it's the board).

As if that weren't enough, a few months ago I bought a Gigabyte GA-880GM-UD2H and had to return it... what was wrong with it? Well, similar to the Asus I mentioned that has me frustrated, that if I modified the memory latencies, if I put them in manual (even if it was to put the same values it picked up in auto), the board wouldn't boot at more than 210-215 bus. The best part is that in my HTPC I have a board of the same model but previous revision (older), and it doesn't have this problem.

On the bright side, recently I also bought another Asus, an M4A88TD-M EVO and another very similar one (also Asus) and with both of them without problems I've raised two Sempron 145 to 3.66 ghz (from the original 2.8) and activated their second core. There they are, both running without problems at 262 mhz bus and with the memory latencies slightly raised to hold 1600@1750 mhz.

So, one of three things, either I'm an idiot and don't know what I'm doing, or the current boards are a lottery and half come defective and aren't good for OC, or I'm very unlucky when buying.

Anyway, I write this post to vent a little and to see if someone can help me

Regards

[elcamaleon]

I think the M5A88 series came out bnnnn messed up. - I have a Pheno 2 x4 965 BE C3… and I couldn't overclock it to 3.8ghz Yle I have a 1.47Vcore which is crazy.--- I tried 2000 ways with multiplier, with bus but nothing... on top of that it raises the temperature like crazy, I changed the power supply thinking it was that, but nothing... now I'm going to buy a discipator, and if it still doesn't let me... I'll throw the Mother in the trash, mine is an M5A88-V EVO.-

[Bm4n]

Well yes, it's getting more and more complicated to do OC, before it was just tweaking 2 things now you have a bunch of values to adjust. And that's apart from the fact that there are boards that leave much to be desired because of the BIOS, or the ones where "auto" simply means I'm going to touch your balls: little laughs:

[Tassadar]

Thanks to both of you for responding.

I want to tell you that last night my balls swelled up (the basketball ones, don't be too quick to judge) and I changed the problematic motherboard (Asus M5A88-M) for the one I said I had with the mutated and overclocked sempron (Asus M4A88TD-M EVO), and without touching anything, leaving everything as it was with the sempron:

-262 mhz bus
-Multiplier 14 and micro to 3666 mhz
-+0,1v micro
-Memory 1,5v@1,64v with slightly increased latencies

It has gone through eight hours of Linpack and today I have left it running the OCCT (the other test it comes with that is not the Linpack). So, leaving the bios as it was, the x6 now works perfectly, and it was just a matter of changing one motherboard for another.

Needless to say, I have tried the configuration that has that motherboard on the other one (this one and a billion more, and stabilizing it not at 262 but at 242 mhz has been completely impossible.

Conclusion? well, either that model of motherboard is simply a fail and is not good for overclocking, or I have been unlucky and got a defective one, because in the same order from pccomponentes was the Gigabyte GA-970A-UD3 AM3+ that I mentioned in the first post, with which you could not mutate the micros (and I hope that with the one they send me as a replacement I can, because if it has the same problem I don't know what I will do).

To be honest, at this point I don't know what to do, if I have to build another computer I won't know which motherboard to mount…....

Regards

[elcamaleon]

@Tassadar:

Thanks to both of you for responding.

I want to tell you that last night my balls swelled up (basketball ones, don't be naughty) and I changed the problematic motherboard (Asus M5A88-M) for the one I said I had a mutated and overclocked Sempron (Asus M4A88TD-M EVO), and without touching anything, leaving everything as it was with the Sempron:

-262 mhz bus
-Multiplier 14 and micro to 3666 mhz
-+0,1v micro
-Memory 1,5v@1,64v with slightly increased latencies

It has passed eight hours of Linpack and today I have left it running OCCT (the other test it brings that is not Linpack). So, leaving the bios as it was, the x6 now works perfectly, and it has been changing one motherboard for another.

Needless to say, I have tried the configuration that has that motherboard on the other one (this one and a billion more, and stabilizing it not at 262 but at 242 mhz has been totally impossible.

Conclusion? well, either that model of motherboard is simply a fail and is not good for overclocking, or I have been given a defective one, and that is bad luck, because in the same order at pccomponentes was the Gigabyte GA-970A-UD3 AM3+ that I mentioned in the first post, with which you could not mutate the micros (and I wonder if with the one they send me as a replacement it will be possible, because if it has the same problem I don't know what I will do).

I honestly don't know what to do at this point, if I have to build another computer I won't know which motherboard to mount…....

Regards

I sadly tell you that yours is not defective… the IA also the one of others also jajjaxD

[Tassadar]

@elcamaleon:

sadly I tell you that it is not defective yours… the ia tmb the others tmbjajjaaxD

You say you have the same model of board and it has the same problem???

Greetings

[elcamaleon]

@Tassadar:

You say you have the same model of motherboard and have the same problem???

Regards

No no … I have the M5A88-V EVO with a Pheno 2 x4 965 BE C3.... and I can't go over 3.8Ghz and MiniMo needs 1.47Vcore to hold that frequency... if I put it at 1.46Vcore blue screen... and I can't go over 3.8Ghz even if you put 1.5Vcore... it's a piece of crap the M5 series... because I put it in a M4A88TD-V EVO and with the stock fan I reached 4.1Ghz with 1.43Vcore hahaha damn.-

[Tassadar]

Well, here we are…....

This whole business with the motherboards is a real mess!

In the end, I gave up and left it with the M4A88TD-M EVO (the one you said worked really well) and I think I'll finally leave it at 3.8 ghz scraped, no more is possible, and I have 1.46 or so of vcore.

I'm going to leave the Asus M5A88-M for a Phenom II x2 @x4 which I'm not going to overclock at all, with the unlocked it's already served xD

Regards and thanks for replying