bytebench

[William Ahern <william_(_at_)_25thandClement_(_dot_)_com>]

I ran bytebench on an Aspen Systems Durango II 633mhz Alpha/512MB RAM I just
purchased from ebay, and on a Celeron 600mhz/128MB RAM. the Celeron seems
to have proved 2X's the performance of the Alpha! How is this? How
can I determine the cache size of the Alpha? Is it on-die? I can't figure
out what the problem is. Is the benchmark just worthless for a comparison?
Included are the results. TIA


=== PIII (Coppermine) 600mhz ===
Arithmetic Test (type = double)               2541.7   176021.2       69.3
Dhrystone 2 without register variables       22366.3  1388801.0       62.1
Execl Throughput Test                           16.5     1747.0      105.9
File Copy  (30 seconds)                        179.0    21587.0      120.6
Pipe-based Context Switching Test             1318.5   105661.9       80.1
Shell scripts (8 concurrent)                     4.0      168.7       42.2
                                                                 =========
     SUM of  6 items                                                 480.1
     AVERAGE                                                          80.0


=== Alpha 21164A-2 633mhz ===
Arithmetic Test (type = double)               2541.7   179184.9       70.5
Dhrystone 2 without register variables       22366.3   711053.4       31.8
Execl Throughput Test                           16.5     1299.7       78.8
File Copy  (30 seconds)                        179.0    13401.0       74.9
Pipe-based Context Switching Test             1318.5        0.0        0.0
Shell scripts (8 concurrent)                     4.0       65.7       16.4
                                                                 =========
     SUM of  6 items                                                 272.3
     AVERAGE                                                          45.4