Hi everyone,
after upgrading the memory of our alpha-workstation to 768MB and
compiling a new kernel with a stacksize of 128MB we seem to be loosing
memory under Dec-OSF4.0D. Adding up the memory sizes of the running
processes we seem to be missing roughly 350 MB running vmstat -P gives
the following output. The size of the ubc-pages seems to be
corresponding with the "missing" memory. Does this seem to be correct
for a cache to grow that large? We would be thankful for any advice.
Henning Loeser, Wolfgang Mannstadt
Total Physical Memory = 768.00 M
= 98304 pages
Physical Memory Clusters:
start_pfn end_pfn type size_pages / size_bytes
0 237 pal 237 / 1.85M
237 98303 os 98066 / 766.14M
98303 98304 pal 1 / 8.00k
Physical Memory Use:
start_pfn end_pfn type size_pages / size_bytes
237 266 unixtable 29 / 232.00k
266 283 scavenge 17 / 136.00k
*
266 280 scavenge 14 / 112.00k
*
283 726 text 443 / 3.46M
726 814 data 88 / 704.00k
814 954 bss 140 / 1.09M
955 1089 kdebug 134 / 1.05M
1089 1094 cfgmgmt 5 / 40.00k
*
1096 1186 unixtable 90 / 720.00k
*
1199 2686 vmtables 1487 / 11.62M
2686 98303 managed 95617 / 747.01M
============================
Total Physical Memory Use: 98064 / 766.12M
Managed Pages Break Down:
free pages = 19923
active pages = 8948
inactive pages = 8727
wired pages = 8101
ubc pages = 49937
==================
Total = 95636
WIRED Pages Break Down:
vm wired pages = 573
ubc wired pages = 0
meta data pages = 2941
malloc pages = 3955
contig pages = 168
user ptepages = 400
kernel ptepages = 56
free ptepages = 6
==================
Total = 8099
Received on Mon Apr 26 1999 - 15:56:47 NZST