- Jul 21, 2014
- 9
- 0
- 4,510
I want a fast build machine for Visual Studio 2013.
On my current machine, the hard-drive bottleneck has been sorted and the next bottleneck appears to be raw CPU horsepower. I have a Q9550 with SSD, and Resource Monitor show all four cores working near capacity and relaxed drive use. It's time to upgrade.
I'm looking at building a new machine with an i7 4930k but wondering if dual Xeon E5-2630 v2's would be a significant gain given the additional cost. It will take the system from 6 to 12 cores, a slightly faster system bus, but 0.8GHz lower clock-speed.
From what I've read, the chips seem similar except the Xeons have integrated PCIe I/O rather than going through a hub, which apparently improves I/O performance by 30%. I plan to put in a PCIe SSD in either case.
What do you good folk reckon?
Also, do you know of any tools out there than can help analyse where the bottlenecks are during a build so I can better target the $'s towards the new machine? I've tried Windows Performance Analyser but didn't find it particularly usable, but maybe that's just me.
Thanks!
Update
I tested out the effect of the 4-cores I have and here are the build times:
1 core: 67s
2 cores: 55s (-12s)
3 cores: 41s (-14s)
4 cores: 36s (-5s)
So adding more cores doesn't look like it will do much. There are a lot of page faults too, so I'll focus my research on a low-latency build.
On my current machine, the hard-drive bottleneck has been sorted and the next bottleneck appears to be raw CPU horsepower. I have a Q9550 with SSD, and Resource Monitor show all four cores working near capacity and relaxed drive use. It's time to upgrade.
I'm looking at building a new machine with an i7 4930k but wondering if dual Xeon E5-2630 v2's would be a significant gain given the additional cost. It will take the system from 6 to 12 cores, a slightly faster system bus, but 0.8GHz lower clock-speed.
From what I've read, the chips seem similar except the Xeons have integrated PCIe I/O rather than going through a hub, which apparently improves I/O performance by 30%. I plan to put in a PCIe SSD in either case.
What do you good folk reckon?
Also, do you know of any tools out there than can help analyse where the bottlenecks are during a build so I can better target the $'s towards the new machine? I've tried Windows Performance Analyser but didn't find it particularly usable, but maybe that's just me.
Thanks!
Update
I tested out the effect of the 4-cores I have and here are the build times:
1 core: 67s
2 cores: 55s (-12s)
3 cores: 41s (-14s)
4 cores: 36s (-5s)
So adding more cores doesn't look like it will do much. There are a lot of page faults too, so I'll focus my research on a low-latency build.
Facebook
Twitter
Instagram