C2D v X2 = Not as good as you think?

Participant · 20 Dec 2006 at 22:22

So are you still trying to prove that C2D is slower at multitasking than X2 by ruling out all possible problems or have you accepted that this is definitely the case and are now trying to reduce the gap for your own benefit?

Nutbusta · 20 Dec 2006 at 22:24

I too would like to setup ACHI mode but my understanding is that this requires a reinstall.

As far as I know, ACHI mode must be enable to use NCQ and/or 3Gbps.

Can anyone confirm this? Also which controller would be better as I'm not running RAID atm?

Sorry for going a bit off-topic

fornowagain · 20 Dec 2006 at 23:02

AHCI does normally needs a full reinstall. It can be done retroactively, you'll have to search for it, involves installing the drivers first, before bios changes. And yes its needed for Sata I & II speeds and NCQ.

FatRakoon · 21 Dec 2006 at 00:47

Robbie G said:
So are you still trying to prove that C2D is slower at multitasking than X2 by ruling out all possible problems or have you accepted that this is definitely the case and are now trying to reduce the gap for your own benefit?

Bit of both, bit of neither.

I dont care anymore about trying to prove anything... I am comparing my own 2 main PCs and I know what I am seeing and I know that my AMD can handle lots more than the conroe, even though the conroe is a much faster CPU, I have run them side by side and I simply find my AMD is much more responsive than the Intel. I dont have to prove anything to anyone and I am still waiting for someone to actually run th very same apps on similar systems to disprove me... Or rather to see for themselves what I mean... While I have had many many people **** me off and saying that their Intel does this or that, not one has actually got back with a head to head where both Pcs are working OTT... Not one.

I would love with all honesty, to suddenly find out that my system is not quite setup right... I really would... If I all of a suddent get the Intels to run the stuff I ask of it as well as the AMDs then I would switch in an instant.

However, I also know that I have it setup as good as I can get it so I dont hold out much hope... Not really.

snarloas1982 · 21 Dec 2006 at 02:18

tried a 5200 x2 against a 6700 conroe both at stock, ran prime, super pi, photoshops cs2 and 3dmark 06. both systems have the same mem (gskill 677 stuff and same gfx 1950xtx) the conroe left it for dust and even finished the 3dmark test minutes before the amd even did.

please stop trying to bum up a slower chip when everyone know the cd2 is faster, i have both and aint fan-boi-istic, catch a grip do you think all the benchmarks from all these test groups are for nothing????

stigggeh · 21 Dec 2006 at 02:30

snarloas1982 said:
tried a 5200 x2 against a 6700 conroe both at stock, ran prime, super pi, photoshops cs2 and 3dmark 06. both systems have the same mem (gskill 677 stuff and same gfx 1950xtx) the conroe left it for dust and even finished the 3dmark test minutes before the amd even did.

please stop trying to bum up a slower chip when everyone know the cd2 is faster, i have both and aint fan-boi-istic, catch a grip do you think all the benchmarks from all these test groups are for nothing????

read the thread...

he is not saying that..... he is saying that running lots of things at once his amd feels faster (i have put in bold the important stuff fot you )

FatRakoon · 21 Dec 2006 at 02:36

Hey hang on... I have not said that the AMD is quicker at any of these tasks...

I said that when I run a fair load of apps, while they both slow down ( obvious ) the Conroe slows down a lot more than the AMD does and it eventually ges to the point, that the AMD can still perform is normal tasks with very little generla slowdowns, while the conroe gets too sluggish.

Ok, so yours is the opposite, and this is what I wanted to know. Mine on the other hand is the opposite to yours. Of course, I can Prime, SuperPI, CS2 ( Although once CS is in, it takes nothing from the CPU so proves nothing ) and sure, 3DMark too, but these for me dont show any real difference betwen the two other than my GFX Cards giving differing scores.

Im not disputing that only running those apps will show any difference to what you had, but try loading a whole load more apps all eating as much CPU % as they can, and you will find that the gap betwen the 2 will close... The most apps you load up the smaller the gap and eventually the AMD will take over.

Or at least thats what is happening to me.

snarloas1982 · 21 Dec 2006 at 02:39

did read the thread..... its just a fan-boi post because his amd chip is not as good as a c2d. where is the evidence that PROVES..... that the amd is faster than a c2d????????????? what is the point of all this jibberish when we need the hard facts for everyone to see? :rolleyes:

Cyber-Mav · 21 Dec 2006 at 02:56

there is something similar to this developing on the xs forums, some sort of server level testing where there is a hell of a lot of stuff going on at once and ram usage averaging above 4gb and dual core opteron system seems to be pulling harder than the conroe setup.

very interesting.

clv101 · 21 Dec 2006 at 07:31

snarloas1982 said:
did read the thread..... its just a fan-boi post because his amd chip is not as good as a c2d. where is the evidence that PROVES..... that the amd is faster than a c2d????????????? what is the point of all this jibberish when we need the hard facts for everyone to see?

What are you on about? This is a discussion, FR is reporting some very interesting experiences - what evidence would you like to see?

Combat squirrel · 21 Dec 2006 at 07:44

*grumpy old man voice*

'Im tellin ye lads its the way the AMD communicates with the rest of the system, its that them there hypertransport and AMD's on bored mem controller, eeeee by gum'

/*grumpy old man voice*

Basically I reckon its defo that, while Intels chip is faster again its the shared cache mem (which makes it faster) which is also its downfall mulitasking, also intel REALLY need to update there FSB tech, its still just a faster version of the fsb used in PC's years ago and its showing its age, it carnt handle the massive amounts of individual data streams going to and from main memory, which is exactly what on die mem controller and hypertransport is designed to do,

a case of AMD slows down less as accessing its ram is direct off die from cpu, hypertransport rest system = good when you run out of cpu cache mem

Intel, while fast still has to go via the chipset on 'normal fsb' tech to ram and back again, this causes lag and therefore the system appears less responsive, while cpu IS faster and has 4mb shared cache, also this system works best when doing 1/2 things at the same time tops, hence why it obliterates running not so much stuff, also im guessing PI is small enough to fit most of it in the CPU's cache mem ?

Any of the CPU's are only as fast as there connection to the ram when there cache mem fills up, so a slower cpu will appear faster as it can fetch data from ram quicker and process it,

What I reckon anyways my 2p aarrrrrr

[HoL]Cobra · 21 Dec 2006 at 08:11

Combat squirrel said:
*grumpy old man voice*

'Im tellin ye lads its the way the AMD communicates with the rest of the system, its that them there hypertransport and AMD's on bored mem controller, eeeee by gum'

/*grumpy old man voice*

Basically I reckon its defo that, while Intels chip is faster again its the shared cache mem (which makes it faster) which is also its downfall mulitasking, also intel REALLY need to update there FSB tech, its still just a faster version of the fsb used in PC's years ago and its showing its age, it carnt handle the massive amounts of individual data streams going to and from main memory, which is exactly what on die mem controller and hypertransport is designed to do,

a case of AMD slows down less as accessing its ram is direct off die from cpu, hypertransport rest system = good when you run out of cpu cache mem

Intel, while fast still has to go via the chipset on 'normal fsb' tech to ram and back again, this causes lag and therefore the system appears less responsive, while cpu IS faster and has 4mb shared cache, also this system works best when doing 1/2 things at the same time tops, hence why it obliterates running not so much stuff, also im guessing PI is small enough to fit most of it in the CPU's cache mem ?

Any of the CPU's are only as fast as there connection to the ram when there cache mem fills up, so a slower cpu will appear faster as it can fetch data from ram quicker and process it,

What I reckon anyways my 2p aarrrrrr

Amen

Hit the nail on the head, 8 pages to get to the conclusion (hopefully)

More or less (more) what I said that a CPU is only as good as it's peripherals or design.

WJA96 · 21 Dec 2006 at 09:04

snarloas1982 said:
did read the thread..... its just a fan-boi post because his amd chip is not as good as a c2d. where is the evidence that PROVES..... that the amd is faster than a c2d????????????? what is the point of all this jibberish when we need the hard facts for everyone to see?

Did you read my post? After 5 days running the same randomly allocated large work units of folding at home running two cores on 2 near-identical PCs the AMD X2 has folded for almost twice as many points as the Core2Duo. That is, in my opinion, incontravertible evidence that under certain circumstances the AMD processors are faster than Intel. I am going to continue to run the two PCs until Saturday morning. Once I have completed it, I will publish the results, listing all the work units and the PPD achieved for each WU by each core on each system. If everyone who posted their opinion was shouted down because he went against the current trend then these forums would be a significantly poorer place. FatRakoon has reported his experience. We are discussing why that varies or agrees with our own experiences and whether it makes a darn bit of difference. If you have something positive to contribute then come on in, we're open minded about your opinion, assuming you have one.

Nutbusta · 21 Dec 2006 at 10:11

WJA96 said:
After 5 days running the same randomly allocated large work units of folding at home running two cores on 2 near-identical PCs the AMD X2 has folded for almost twice as many points as the Core2Duo. That is, in my opinion, incontravertible evidence that under certain circumstances the AMD processors are faster than Intel.

That may be so, but it's hardly a fair test is it? What you didn't mention is that you've been ripping you dvd collection and playing NFS: Carbon on the conroe 24/7

WJA96 · 21 Dec 2006 at 10:31

Nutbusta said:
That may be so, but it's hardly a fair test is it? What you didn't mention is that you've been ripping you dvd collection and playing NFS: Carbon on the conroe 24/7

Sadly not. In this particular case they are just doing 3 things each. 2 threads of F@H 24/7 and logging the CPU utilization of each core.

Once in a while I will just log in over an IP KVM and check they are still running away fine, but other than that, that's all they're doing. I take my folding very seriously so this is quite important to me.

It just happens to tie in with FatRakoon's observations so I thought I'd post it. In the same way that Athlon 64's were not always faster than Pentium 4's I would expect there to be some circumstances where AMD X2's will outrun Core2Duo's. As most of these large WU were optimised for the Pentium 4 the initial expectation might be that they should run better on later Intel systems however as many people have pointed out the Core2Duo architecture draws more on than Pentium Pro than anything since then, so maybe Folding@Home isn't optimised for Core2Duo and because of that the AMDs are doing better in this case.

Whatever the reason, I'm going to optimise my folding systems because they are actually quite expensive to run, both in terms of money and carbon emissions.

Rizz · 21 Dec 2006 at 14:28

carbon emmisons ?

Rizz · 21 Dec 2006 at 14:29

carbon emmissions ?

clv101 · 21 Dec 2006 at 14:52

Rizz said:
carbon emmissions ?

The social cost of electricity consumption. The 'rational' individual maximising personal utility will disregard it however anyone whose values extend beyond the personal is correct to consider such social costs.

Rather that just “carbon emissions” it would be more complete to consider the dead miners, the increased number of respiratory problem sufferers, currently non-monetised costs of nuclear decommission and gas depletion etc – basically the costs of marginal electricity that aren’t covered by the 10 pence per kWh you pay as a private cost.

WJA96 · 21 Dec 2006 at 15:16

clv101 said:
The social cost of electricity consumption. The 'rational' individual maximising personal utility will disregard it however anyone whose values extend beyond the personal is correct to consider such social costs.

Rather that just “carbon emissions” it would be more complete to consider the dead miners, the increased number of respiratory problem sufferers, currently non-monetised costs of nuclear decommission and gas depletion etc – basically the costs of marginal electricity that aren’t covered by the 10 pence per kWh you pay as a private cost.

Furby vRS owners of the world unite. You have nothing to lose but your stutter!

bfar · 21 Dec 2006 at 15:28

Don't see what all the fuss is about. The two CPUs have very different architectures, and it stands to reason they'll have different strenghts. Some more than others. If the poster sees better performance for his needs on AMD, then thats great

I'll swap over to C2D in the new year because I play games more than anything else. If I was just Folding all day I'd stick with my X2 4800+. And why not if thats what you do?

I upgraded to X24800+ because at the time I just needed to buy the CPU. Nice and easy, and just €300. When I buy my new setup I'll need a mobo and memory - €500 to €800 depending.

DIFFERENT FOLKS, DIFFERENT STROKES. Don't stress over it.