User panel stuff on forum
  8 posts on 1 page  1
Client Talk
2009-01-28, 06:53
Member
100 posts

Registered:
Apr 2008
noticed changing the values leads to a change in performance/feel of the game. even removing a few extra processes in task manager seems to assist.

what do people run as it seems quite different running -heapsize 131072 for 128 meg or 655360 for 512.

is there a consensus on what should be run is there a slowdown eventually if you increase too much?

what is the deal with this command and what should be run. running a fair beast atm btw with 4gig ram so can probably run anything up to a gig or so. any feedback would be appreciated
2009-01-28, 09:24
Member
1100 posts

Registered:
Jan 2006
If you are using ezQuake you shall not tinker around with that. The engine does not need it.
2009-01-28, 09:53
Member
100 posts

Registered:
Apr 2008
yes but there is a noticable difference between different allocations of heapsize. so allocating no memory is best? is that what your saying?
2009-01-28, 11:57
Member
271 posts

Registered:
Feb 2006
if you allocate too much heap, the spare pages shouldn't be used (they're not even mapped in FTE).
Quite frankly, if heap size makes a noticeable difference on a 4gb machine then something really fucked up is happening, and you should try installing linux instead...

Quite frankly, you shouldn't need more than 16mb, and more than 32mb for a 32 player deathmatch map would have to be a shockingly detailed map.
moo
2009-01-28, 16:17
Member
100 posts

Registered:
Apr 2008
yeh i think i realised it just sort of makes it smoother in a way but it seems slower so yeh dont think you need it
2009-04-17, 10:33
Administrator
384 posts

Registered:
Dec 2006
I discussed this with fuh in the pre-ezquake days and his opinion was that there was no value from having a heapsize larger than 32MB.

I'd be interested to know peoples view on democache however. I've run with 33333K for years (seemed to be plenty for a 30min match) but I wondered if it needed to larger now we're running with much higher framerates.... I'm guessing it's only physfps (typically 77) which affects demosize when you have async client framerate, or does that affect demosize too?
2009-04-17, 16:03
Member
485 posts

Registered:
Feb 2006
HangTime wrote:
I'd be interested to know peoples view on democache however. I've run with 33333K for years (seemed to be plenty for a 30min match) but I wondered if it needed to larger now we're running with much higher framerates.... I'm guessing it's only physfps (typically 77) which affects demosize when you have async client framerate, or does that affect demosize too?

Only physframes get recorded. I accidentally removed democache and never noticed any difference. Had it at 100M, because why not. TF demos were over 60M sometimes.
2009-04-18, 05:02
Member
252 posts

Registered:
Dec 2006
krohm wrote:
yeh i think i realised it just sort of makes it smoother in a way but it seems slower so yeh dont think you need it

BAHAHAH *scoff*
'on 120 ping i have beaten mortuary dirtbox and reload' (tm) mz adrenalin
'i watched sting once very boring and not good at all' (tm) mz adrenalin
[i]'i shoulda won all
  8 posts on 1 page  1