| #36. Posted at 11:09 PM on Nov 2nd 2007 | Edit Reply |
|
Shining Arcanine |
Did Intel not do this with the Itanium 2 a while ago?
 |
| Edit Reply |
|
flip-mode |
My initial reaction to this is that it's not very exciting. I'm ready to be surprised though.
|
| Edit Reply |
|
JJCDAD |
But how do you know which core an app is going to run on? Sometimes it runs on the fast core, sometimes on the slow one? That doesn't sound fun.
Multi-core apps running on cores that are running at different speeds? Something's gonna splode? We'll see. |
| Edit Reply |
|
alex666 |
Is this a gimmick, or might this genuinely have some utility, as well as being an example of AMD thinking outside the box?
|
| Edit Reply |
|
albundy |
and if you do overclock each core, how would you really know that the o/c is successful?
|
| Edit Reply |
|
Usacomp2k3 |
You could do dynamic overclocking ie have 2 cores run at 50mhz and 2 run at 3ghz, and then every 1 minutes switch them. That should help keep things from overheating.
|
| Edit Reply |
|
ssidbroadcast |
Well, okay. I'd sign off on this. Might not be a well-used feature but OC-ers/enthusiasts will like it.
|
| Edit Reply |
|
insulin_junkie72 |
Heh, if you take a look at the screenshot in the expreview article, the PCI-E speed sure takes a hit with the core OCing :P
|
| Edit Reply |
|
wingless |
We know this about Phenom X4's now: AT LEAST ONE OF FOUR CORES CAN OVERCLOCK TO AT LEAST 3.33GHZ. We can be pretty damn sure thats not the max overclock for the final chip. I'd put my money on that. I have to emphasize again just how much I can't wait to see the final performance of a real Phenom X4. Paired with the upcoming RD790 we should see some encouraging results.
|
| Edit Reply |
|
computron9000 |
That would be quite impressive. It would certainly make signatures more complex on tech forums.
"AMD Phenom 3.1Ghz/3.2Ghz/3.44Ghz/2.8Ghz" I wonder if avg. Ghz will become the new way to refer to your processor. Like, "AMD Phenom x4 ~3.22Ghz". It will, however, make over clocking considerably more time consuming to achieve stability. There would be 12 different approaches in a "one by one" core overclocking approach. i.e. o/c one core, test, o/c next core, test, etc. It would also be interesting in how heat and voltages correlate to application use when certain cores are over clocked, since it would not be identical for each core because of which programs utilize what cores at what capacity. Given how multi-threading is handled in a dual-core system, I suspect differently clocked cores in a quad-core system, using a program running 4-threads would all achieve different utilizations and impact the overall heat/stability differently. |
| Edit Reply |
|
DASQ |
It would definitely be better for getting the best overclocks out of a given CPU, as one core lagging behind no longer holds back the other three.
As long as each core has their own individual voltage settings, as well. |