or Connect
New Posts  All Forums:

Posts by Anonymous Karma

I still do not understand why the mac web went absolutely NUTS here and for no reason whatsoever decided to start questioning whether or not the iBook was a "real" G4. Was it that surprising? Did people not believe that MacRumors, AppleInsider and NMR could be wrong? Well, I'm sorry folks, but if the rumor sites and Apple say two different things about an announced product, Apple is right. So, tell me, all of you that decided to jump on this paranoid bandwagon: why? Did...
Bzzzzzzzt! The G5 uses HyperTransport for connecting the North bridge to - well, to the I/O controllers to the south of it. The elastic bus that connects the two CPUs to the system controller is /not/ HyperTransport. Care to try again?
Seems quite provable to me that the defined factorial function has no side effects and will terminate when called with the value 16... thus the call to factorial can be replaced with the static value, and then the whole loop optimized out. Or perhaps I'm just used to saner languages than C?
Edited: Whoops! That's what I get for not reading the thread. As Programmer said most of this test is function call overhead. And any good optimizer will just optimize out the whole loop
Actually a dual Opteron just has two memory controllers. As I said for embarrassingly parallel work the Opteron is nice, but the reason most people SMP and not a cluster of $300 Linux machines is because most problems are not /that/ parallel. Across the bus between the two processors there is only 6.4GB/s and this bandwidth, unfortunately enough, will be conflicted between memory transfers and MOESI coherency transactions. I'm sure there are other applications where the...
Aaaagh! Pet peeve! I think you meant "try to do". Anyway, congrats!
Uh, no. The problem is that the Athlon 64 / Opteron are NUMA systems, which work great for massive scalability but run into some bottlenecks on a 1-processor system - and some worse issues on 2P systems. One of the goals of system designers over the past ten years has been to allow each of the various components of the system to speak to each other with a minimum of bandwidth competition. Now, when looking at a buzzword list for the Athlon 64, you'll find a lot of terms...
Mathematica has not yet been optimized for the G5 (or rather it has not yet been released in optimized form). The autovectorization capacity of IBM's compiler is still limited. The rest of those numbers are even more meaningless as you can't tell what they're doing. Note they did not check the Opteron on the other benchmarks either.
Nope, and worse they've been complete jerks to us over the phone too. (Still awaiting a CTO G5 - and the only thing I did was put in the 250GB. Stupid, stupid, stupid, stupid, ...)
Mike Z sucks. He's made negligible progress with the cell phone division and is a Galvin lackey. Let's hope the board brings in someone from the outside who is willing to shake up the Old Boys' Club.
New Posts  All Forums: