[distcc] call for testing

Burt Holzman burt at bnl.gov
Thu Jun 24 05:08:17 GMT 2004


Here's results on our mini-cluster (22 dual nodes, the faster half are 
1000 Mbit, the slower 100 Mbit) compiling our analysis codebase.
It's a production cluster, so I couldn't run any detailed tests.

LZO with MMAP 
170.060u 100.330s 3:59.45 112.9% 0+0k 0+0io 1589626pf+0w
179.970u 94.590s 3:35.29 127.5% 0+0k 0+0io 1584071pf+0w

LZO without MMAP
180.640u 96.070s 3:14.13 142.5% 0+0k 0+0io 1549227pf+0w
177.780u 95.560s 3:14.95 140.2% 0+0k 0+0io 1554611pf+0w

No LZO with mmap
166.860u 93.030s 3:20.19 129.8% 0+0k 0+0io 1582927pf+0w
169.700u 92.450s 3:31.66 123.8% 0+0k 0+0io 1583767pf+0w

No LZO without MMAP
178.480u 93.490s 3:14.14 140.0% 0+0k 0+0io 1546516pf+0w
207.090u 93.400s 3:21.60 149.0% 0+0k 0+0io 1550266pf+0w

- B

More information about the distcc mailing list