Are dual processor machines necessary ?

Quote from BobbyMurcerFan:

Xenons have more cache and had hyperthreading before the P4 did. There is really no reason, IMO, to have a Xenon machine. You are looking at a much more expensive CPU and motherboard when compared to straight P4 of similar clock and FSB speed. And the Xenons always lag the P4 in terms of clock and FSB speed. For example I believe the fastest Xenon available is 2.2 gig w/ a 400 Mhz FSB. While the fastest P4 is 3.06 gig w/ a 533 Mhz FSB soon to be (as in any day now) 800 Mhz. I think a single top of the line P4 will outperform a single top of the line Xenon.
Actually, the fastest Xenons are 3.06 gig w/ 533 Mhz FSBs. And the latest P4's have 512K of L2 cache, so the line between Xenon and P4 seems to blur more and more every day.
 
Quote from ArchAngel:


Apparently not - didn't Nitro say he routinely runs his dual config at 85% all day, every day?

Yes, but I also heard several other things that would give me rise and understanding as to some of the issues there. I'd querry the setup of the hardware just a bit harder. :)
 
Nitro, I guess you wrote your own software for scanning, could you elaborate a bit on how you did this ? Isn't the problem with scanning options not also with what you put into your models ? (Vol, dividends, skew ...)
 
quote from canyonman00:

Yes, but I also heard several other things that would give me rise and understanding as to some of the issues there. I'd querry the setup of the hardware just a bit harder.
If the machine is running hard for that long, there are some software issues.

Normally the activity of the computer should correspond to the activity of the market. When the market is active and volatile your machine may get as high as 50 percent, not much more. But when the market is dull and listless your machine should be a lot lower than 50 percent.
:cool:
 
Quote from spreadem:

If the machine is running hard for that long, there are some software issues.

Normally the activity of the computer should correspond to the activity of the market. When the market is active and volatile your machine may get as high as 50 percent, not much more. But when the market is dull and listless your machine should be a lot lower than 50 percent.
:cool:

Exactly! So a pegged resource usage exposes another flaw of some type. But there could also be many more. A slower spinning hard drive (5400rpm vs. 7200rpm vs. 10,000rpm) for example can cause a special lag of its own too! :)
 
A steady, high CPU usage doesn't necessarily mean there is a software problem (although it could of course). If a program is being fed data fast enough, then the CPU can maintain a constant level. So all this is speculation as we don't have access to the programs and source code he is using, etc.
 
Quote from canyonman00:
Yes, but I also heard several other things that would give me rise and understanding as to some of the issues there. I'd querry the setup of the hardware just a bit harder.
Quote from spreadem:
If the machine is running hard for that long, there are some software issues.

Normally the activity of the computer should correspond to the activity of the market. When the market is active and volatile your machine may get as high as 50 percent, not much more. But when the market is dull and listless your machine should be a lot lower than 50 percent.
Wow, you guys are amazing. To be able to offer such decisive expert conclusions without knowing anything about the specifics of the workload, configuration, detailed utilization stats, etc. is far better than most people could do :)
 
Back
Top