Page 1 of 2

50 000 units

Posted: 19 Jul 2006, 11:41
by emmanuel
someome prob by s3o framerate can be huge lowering
this get me to think to re lag this by increasing unit count...
isn t?
a unit limit can be edit in files??? or engine coded

Posted: 20 Jul 2006, 03:06
by Neddie
I would like a higher maximum unit cap, for future computer game usage.

Posted: 20 Jul 2006, 03:44
by emmanuel
but for now a modder using format .s3o
and noanimated low poly unit
hav report a reduced frame rate
and then posibility increasing the unit number

is only whise a amd2200 that ic an play at totalA with 1500units(never with5000)
i hav buy the first release for a pentium100 in 1994 maybe

Posted: 20 Jul 2006, 03:47
by SwiftSpear
The majority of spring lag is caused by patching information being stored and worked on for each individual unit. Spring makes the most work for your CPU, with enough units even the strongest CPU's slow down.

A High end graphics card should be able to run spring with 50000 units, but the CPU will not last long enough.

Posted: 20 Jul 2006, 03:52
by emmanuel
with 2Gb ram and a 4GHz dual core
how many unit can be used?
my not64bit 2800+ can suport easily 2000units

...about 64bit :is the taspring must be improved for exploit this?
and for dual core?

Posted: 20 Jul 2006, 04:22
by SwiftSpear
Spring doesn't use any 64 bit calculations. If we could use 64 bit native calculations it wouldn't make any difference almost at all most of the time. The 64 bit processors have better optimizations then the 32 bit AMD's so they are still worth getting, but they don't use their 64 bit capability in spring.

I'm not sure how many units a 2Gb ram and 4GHz dual core machine could run, it should be quite a few though...

Posted: 20 Jul 2006, 06:32
by KDR_11k
The 64 bit procs have 8 additional registers which is where the speed boost comes from.

Dual code shouldn't make a difference, I don't think Spring is multithreaded.

Posted: 20 Jul 2006, 08:58
by emmanuel
is 32bit to 64bit with corect code can 2xframerate?
is a multithread code can 2x framerate again ?
2 x 2= 4 x
maybe the ended realese must hav all modernity....

but i always nounderstand why taspring don t work in win98: what is new in winxp ?

Posted: 20 Jul 2006, 18:35
by Drone_Fragger
emmanuel wrote:is 32bit to 64bit with corect code can 2xframerate?
is a multithread code can 2x framerate again ?
2 x 2= 4 x
maybe the ended realese must hav all modernity....

but i always nounderstand why taspring don t work in win98: what is new in winxp ?
64 bit is really just a waste of time. So few games are optimised for it that it barely improves framerates. Although a 50,000 unit limit would be nice.

Posted: 20 Jul 2006, 18:46
by emmanuel
to 50000 unit it reach the count to real historic battle ...in middle age ...

Posted: 20 Jul 2006, 22:44
by jcnossen
Spring already runs in 64 bit on linux, but not with any spectacular speed bonus I think.

Posted: 21 Jul 2006, 07:38
by Das Bruce
emmanuel wrote:but i always nounderstand why taspring don t work in win98: what is new in winxp ?
Completely different base, XP is based of NT.

Posted: 21 Jul 2006, 15:14
by KDR_11k
emmanuel wrote:is 32bit to 64bit with corect code can 2xframerate?
No.
is a multithread code can 2x framerate again ?
No.

Posted: 23 Jul 2006, 13:51
by emmanuel
so the story is revers=
last time win nt hav no games...

don t understand why 64bit or multithread is not helpfull for improv code???

can you will be explicativ in your answer:no yes maybee

Posted: 23 Jul 2006, 14:09
by SwiftSpear
Well, in order for 64 bit processing to improve the code the code needs to have alot of 64 bit variables used. 64 bit variables are rarely needed in code, so most coders only put very few of them in their programs, another reason is that they are one of the things that slow down 32processing, so generally they are avoided for programs that want to be efficient for the widest margin of users. 64bit can improve calculation efficiency for specially written programs, but rarely helps much for programs that aren't written to use it. There's not much point in rewriting spring to use 64 bit because it would be less usable for most users, and we probably don't have alot of area's where 64 bit variables would help us much anyways.

Posted: 23 Jul 2006, 14:22
by emmanuel
when i ask :"what for the spring don t work in win98?"
it answer "don t care about minority" :
32bit are minority too...

Posted: 23 Jul 2006, 16:42
by Hunter0000
emmanuel wrote:when i ask :"what for the spring don t work in win98?"
it answer "don t care about minority" :
32bit are minority too...
Win 98 is not aminority. Its an old, unsafe, abandoned OS that should not be touched with a pole.

Very, Very different.

As other have stated, rewritting parts of spring to use 64-bit would likely slow down more users then it would speed up.. so there really is not much point...

Posted: 23 Jul 2006, 17:55
by Sgt Doom
Hunter0000 wrote:Win 98 is not a minority. Its an old, unsafe, abandoned OS that should not be touched with a pole.
That and it crashes more than a drunk driver on crack and weed.

Posted: 23 Jul 2006, 18:18
by Guessmyname
Sgt Doom wrote:
Hunter0000 wrote:Win 98 is not a minority. Its an old, unsafe, abandoned OS that should not be touched with a pole.
That and it crashes more than a drunk driver on crack and weed.
Probably because drunk drivers on crack and weed can only crash once before either:

A, dying

or

B, getting arrested

Posted: 23 Jul 2006, 21:11
by KDR_11k
Both 64bit and multithreading for multiprocessor systems would give performance boosts on the appropriate systems but it wouldn't be even close to double. 64bits isn't so much about the bits as it is about additional registers which the IA32 desperately needed. Multithreading isn't perfectly efficient and causes a lot of overhead.