Overclocked Opteron vs. Fx-57 and others

[n91htmare]

I was sitting here pondering about this and I have questions.
I picked up a opty 146 on which I was able to hit 3gigs on stock air (idle temp at 44) on MSI Neo plat 2. This was fairly an easy oc.

Now, I see the FX-57 and other chips like the 3700+ and 4000+ and so on.
Now the opty's clock speed is faster than all others at stock. But lets just take the opty and FX-57 for now.

What is the difference of an opty at 2.8 and FX-57 at 2.8? both have the same amount of L1 and l2 cache, is the Quality of the Cache and Core? Would the quality improve processor speed or will it just add durability and stability in the long run?

Also, what would be the opty at 3gigs be relative to? like a 4800+ or something?

just pondering and learning.

[`SippY]

the fx-57 is faster becasue of the better memory controller

[CrashOv3r1De]

the fx-57 is faster becasue of the better memory controller

Agreed. You have to clock the opteron a decent amount past the FX57 to beat it

[zir_blazer]

He is probabily talking about Opterons 1xx for Socket 939, those uses exactly the same San Diego or Toledo Cores that any other Athlon 64 and gives the same performance at the same Frecuency. There are no changes in the Memory Controller, amount of Hyper Transport Links or anything because they uses the same Cores.
The only diference is when overclocking, because those Opterons uses high quality Cores (Some of them with A64 FX-57 Stepping) and seems to have a more higher overclocking capabilities.

[k00lance]

aside some features, and both cpu @ same speed they are about the same..
The reason you might want fx is because it has more potential such as unlocked multiplier.

[s7e9h3n]

What is the difference of an opty at 2.8 and FX-57 at 2.8?

~$800

[k00lance]

I think until opt s939 came out.. fx-57 served for really good purposes..
For air users able to reach high clocks without sweating..
and for people who like to freeze their chip

[Cooper]

I think Opty on 2.8GHz will perform slightly better than stock FX-57 cos o` higher HT speed

[Vapor]

Stephen is right on there....they're not the exact same core (the registers for the name and stock settings are different ) but if you clock them to the same speed and settings, they'll perform the same.

And the unlocked multi crap is just that--crap. You can get any CPU:memory ratio that's worth while you want with clever uses of various multis and dividers.

Of note is that the FX is guaranteed to run at 2.8GHz at nominal voltage--not all Optys can do that

[s7e9h3n]

Well after reading a few replies here, maybe I should chime in with something a little more useful :
Vapor's right - 2 cpu's set with the exact same timings and using the same hardware should theoretically perform the same. BUT here we're talking about the difference between the opty 146 and the FX57........Although I've never tested a 939 Opty, I hear there are issues with certain memory dividers and settings? One of those issues concerns the MAL an RP on the s939 Opties. Seems they have to run abnormally high values for those settings. That in itself would make a difference in benchmarks as every FX57 I've seen has no problem running the tightest memory settings available. This is my 57.....show me an Opty that can run anything close to these timings -

[ben805]

someone did a SPI 1m, 8m, and 32m comparing the opty and fx57 at the exact same clock speed and same timing, the result is virtually identical. that post is somewhere around here so look around, though I still think the memory controller of the fx is superior than the opty as mentioned by Stephen, other than that and the unlock multiplier of fx, the only other difference is price, value, and bragging right

[s7e9h3n]

someone did a SPI 1m, 8m, and 32m comparing the opty and fx57 at the exact same clock speed and same timing, the result is virtually identical......

This should be expected as they share the same core and memory controller architecture. But out of curiousity, was the Opteron able to run the same ram timings as the FX? Or was the FX basically "sandbagging" it's memory timings?

[n91htmare]

thanks for the replies.
I suppose the following could also be put in perspective.
Where the Opteron reaches it's maximum (2.8-3.0), the FX-57 is just getting started.

Now I'm not too sure if the $840 difference is worth it.

[dogsx2]

Well after reading a few replies here, maybe I should chime in with something a little more useful :
Vapor's right - 2 cpu's set with the exact same timings and using the same hardware should theoretically perform the same. BUT here we're talking about the difference between the opty 146 and the FX57........Although I've never tested a 939 Opty, I hear there are issues with certain memory dividers and settings? One of those issues concerns the MAL an RP on the s939 Opties. Seems they have to run abnormally high values for those settings. That in itself would make a difference in benchmarks as every FX57 I've seen has no problem running the tightest memory settings available. This is my 57.....show me an Opty that can run anything close to these timings -

Show me 10 more FX57's that can do that.

[DeToNaToR.cl]

i had seen some optys getting 23s en 1m

Is 1 second worth 820usd?

[tatts]

it might be 3.6ghz which a few opties have done with the same cooling but what speed is your ram running to get them timings & what ram is it? coz i dont think its down to the cpu

[gundamit]

i had seen some optys getting 23s en 1m

Is 1 second worth 820usd?

There's no correct answer. For some wanting to having the absolute best and leave very little to chance, it might be worth it. To others 820usd is a lot of beer money.

[ben805]

This should be expected as they share the same core and memory controller architecture. But out of curiousity, was the Opteron able to run the same ram timings as the FX? Or was the FX basically "sandbagging" it's memory timings?

I found that thread of Opty vs FX, here

both Opteron and FX were running at the exact same timing.

[Vapor]

Stephen....I think only the original CABNEs needed the higher RP and MAL and it definitely slowed them down. The odd thing was that with cold temps the changed RP and MAL weren't needed and that FXs with identical steppings didn't gain from the different RP and MAL. The newer ones don't seem to clock as well on average but also don't seem to need the altered timings.

I personally had ZERO problems with the dividers and multis (thought I did at first but realized I was doing something dumb) on my CABNE0530APMW and could get any CPU:RAM ratio I needed.

On average, the FX is the better way to go because it never has a problem reaching the high speeds or sticking with the tight timings, but it seems Optys are beginning to hold their own in those respects as well--especially with colder temps.

Also of note is how hot these Optys get--that was very curious to me. I have seen zero reasons WHY though (I still think there's a bit more difference between Venus and San Diego than the name and clock info registers, but no definitive evidence from AMD yet).

[Vapor]

I found that thread of Opty vs FX, here

both Opteron and FX were running at the exact same timing.

Yes, but was the FX going as tight as it could or were the timings dumbed down to the Opty's timings?

[zir_blazer]

I found that thread of Opty vs FX, here

both Opteron and FX were running at the exact same timing.

And that is the proof that they're identical with the exeption of nominal specifications and name.

Remember that as every Processor overclocks diferently, the Memory Controllers can also react diferently (Not that there is any technical diference bewthem them), the only that you can get guaranteed is what the nominal specifications that it tells you.

[tatts]

the only difference to me is you have more money than sense to go for the fx or wise if you go for the opteron

[Theo404]

every Processor overclocks diferently, the Memory Controllers can also react diferently

Couldnt agree more. Until an insider can tell us that they test FX57 memory controller capabilities to a greater extent that they do opterons, then its just each chips individuality...... And even if they did employ a different memory controller performance test when binning, wouldnt you think that they would test busisness grade chips just as much (if not more) than flagship chips?

[dogsx2]

the only difference to me is you have more money than sense to go for the fx or wise if you go for the opteron

Money only does you good if you spend it. If you have lots of money and do not spend it, what is the difference between you and the man who has no money? Neither of you bought so you are no better off then he is.

[Theo404]

Money only does you good if you spend it. If you have lots of money and do not spend it, what is the difference between you and the man who has no money? Neither of you bought so you are no better off then he is.

So you think people should spend money just because they have it?????

In your example, the person with money is better off compared to the person will no money, he may have the same hardware as the poor guy, but hes still got money in his pocket to spend properly.....

Unless you have huge amounts of disposable income, or your LIFE is benchmarks, buying an FX is ridicules.