PDA

View Full Version : HL2 map compile overheats my CPU



Biggles
10-03-2005, 12:05 PM
Since I ditched my cruddy old case in favour of a Li-an Li VC1000 a while back, I've had no overheating problems, even during the hotest days. But now that I'm compiling CS:S maps I've got a new heat issue. CPU (P4 2.6GHz) heat rises rapidly to about 68 degrees during the final (RAD) process in my map compile. And this is regardless of ambient room temp (happens at night).

Must be some serious number crunching going on there. Since it appears this is directly related to the CPU working - and not related to any video card or hard drive generated heat - it looks like I might need to investiagte some better CPU cooling options to replace the standard Intel heatsink/fan combo that on there at the moment.

Anyone got any recommended CPU/fan combos for a P4 2.6GHz?

Chilling_Silence
10-03-2005, 12:55 PM
Ive always been a fan of the Volcano series of fans :D

Not too expensive either, I picked up a Volcano9 about 6 months back for only $40ish

What happens if you do some benchmarking and max out the CPU a bit, what happens to the temp then?

ninja
10-03-2005, 01:07 PM
Go AMD. Feel the love.

Biggles
10-03-2005, 01:13 PM
Go AMD. Feel the love

I wouldn't mind trying except ...... $$$$$$$$$ ........ major upgrades are not an option right now, and when an upgrade does poke over the budget horizon, it will likely be the vid card.

Lohsing
10-03-2005, 01:21 PM
Hi Bruce - assuming you have a Prescott Socket 478 CPU there?

I've heard great things about the Zalman CNPS7000 (http://www.endpcnoise.com/cgi-bin/e/p4_cnps7000-alcu.html) as a cooler.

Lo.

Biggles
10-03-2005, 01:27 PM
What happens if you do some benchmarking and max out the CPU a bit, what happens to the temp then?

Haven't tried. Might take 3DMark home and give it a go. There's no video load involved in this - map compiling is running in a command box and no other apps are running at the time.

Murray P
10-03-2005, 01:37 PM
Once you have established for sure that it's the software load on the CPU rising your temperature, have a wander around here (http://www.silentpcreview.com/section10.html) to get the lowdown on whats hot & whats not, without driving the rest of the household barmy while your compiling.

PaulD
10-03-2005, 01:51 PM
What is the CPU fan speed at the time? The stock fan is temp controlled, perhaps it takes too long to ramp up the speed.

Chilling_Silence
10-03-2005, 01:54 PM
Yes, but the benchmarking should also put pressure on your CPU, unless you feel like booting the INSERT Linux LiveCD and running "stress" which will load your CPU

Lohsing
10-03-2005, 01:58 PM
There's always the SiSoft Sandra "burn in" function...

Lo.

Biggles
10-03-2005, 02:06 PM
Muhahahahaha I like the look of this:

http://www.silentpcreview.com/article186-page1.html

Murray P
10-03-2005, 02:19 PM
Got your HT license Bruce?

Biggles
10-03-2005, 02:58 PM
Hahahahahahahahahah - Now we're talking:

http://www.viperlair.com/reviews/cooling/misc/asetek/vpls/

Metla
11-03-2005, 09:46 PM
Send me that Lan-li Bruce and I'll give it a custom chop, Similar to the job I done on mine, Though these are true custom jobs, No two are the same.

http://sal.neoburn.net/imagef1/files/toddlertamer.jpg

Can you believe I managed to do that with only an electric drill and a set of pliers?

Amazing I know.

tweak'e
11-03-2005, 10:49 PM
bruce....the standard heatsink should do it ok.

whats your case temp like? and does it go up when the cpu overheats?

Myth
11-03-2005, 11:04 PM
Ever considered water cooling?

Hmm Bruce Buckman, subject of the next Pimp my PC :cool:

Biggles
15-03-2005, 01:23 PM
The heat spike appears pretty localised. Unlike my heat problems last year, where case temp went noticeably up, with this particular issue the case temp only rises a little bit, but the CPU temp rockets from an idle temp of 39-41 up to 68-69 in a few minutes.

I've come to the conclusion that I must have messed up the installation of the heatsink/fan when I built the PC a while back. As Tweak says, the standard unit should be adequate. It looks to me like its functioning at only moderate efficency and when the load really goes on, it loses the plot big time.

But never one to let an opportunity pass I've decided to replace it anyhow with one of those Thermalright Xp-120s and use some Artic Silver 5 compound for no other reason than I want to see what happens :-)

And while I might do a water-cooled job one day, it won't be on my PC. I still can't get my head round the concept of mixing water with electronics .....

Biggles
15-03-2005, 01:34 PM
Can you believe I managed to do that with only an electric drill and a set of pliers?

Looks like what I had to do when my cat fell down behind our pantry trying to catch a dumb bird (which had already fallen down behind the panty).

The whole unit was sealed except (obviously) from above and the only way I could get the cat out was to take a drill to the inside, punching enough holes through the customwood to allow me to knock out a "window" for the cat to crawl through.

My workmanship was every bit as precise as yours ....

Murray P
15-03-2005, 02:13 PM
VapoChill Micro (http://www.asetek.com/default.asp?showPage=startside.asp&param=sideid&myvalue=14&contentSection=2&menuID=-1)

Biggles
16-03-2005, 01:07 PM
Darn, the Xp-120 won't fit in my case. Never mind.

But I've taken the heatsink off, cleaned it and the CPU off with the old isopropyl alcohol, and then reset it using some Artic Silver 5. It's going to take a while to thoroughly burn in but initial results are encouraging - doesn't peak above 56, over a 10 degree drop at this point. I'll have to see how it is a few days from now.

E|im
16-03-2005, 05:11 PM
But I've taken the heatsink off, cleaned it and the CPU off with the old isopropyl alcohol, and then reset it using some Artic Silver 5. It's going to take a while to thoroughly burn in but initial results are encouraging - doesn't peak above 56, over a 10 degree drop at this point. I'll have to see how it is a few days from now.
That's a good method. What are you using to burn it in? I would suggest Prime 95 torture test (Max heat, power consumption one). If your P4 has hyper-threading run two instances, if not just one instance. Run it for like an hour.

Should put it to the real test. :)

Biggles
16-03-2005, 08:26 PM
Artic Silver say normal usuage should be enough over a 200 hour period, as long as it involves shutting off and letting the PC return to normal temp every so often.

The map compile process is a pretty good burn in by itself. CPU utilisation is basically 100% while its going - 10-15 minutes for my current project depending on compile options used - and I compile roughly twice and hour while I'm working on it.

In fact, I'm going to compile right now ...

Metla
16-03-2005, 08:41 PM
Just to be nosy,

Can i ask why your compling maps?,presumbly you have created a map for some specif reason,or edited an exsisting one?

Do you share them around,why mod them at all?

Biggles
16-03-2005, 11:35 PM
Not modding em boyo, I make em. Made my first map for Doom 2. I've mapped for Quake, HL1 deathmatch, CS, DoD and now I'm doing CS:S maps while I wait for DoD:Source to arrive, which is my HL mod of choice.

Started making maps for our internal lunchtime deathmatch sessions years and years ago. Made an early Doom2 map of our office space once where I hid assorted power ups in a secret room in my "office". Used to wax Chris Keall's butt till he caught me opening the secret room up and cried foul.

I made a series of quick and nasty CS maps back before there were decent bots available. Made em small so 6 to 8 players could have fun which was about how many people we had playing at our peak.

Didn't publically release any of these because they - well quick and dirty. But I did release a pretty good DoD map. Had flaws and I'm going to remake it for DoD:Source. It worked well with Sturmbots though.

Metla
16-03-2005, 11:42 PM
Excellent, the secret stash of weapons is a work of genius,or extreme cunning anyway.

As dor DoD, I've never played it,know nothing about it except its a ww2 setting,Does it follow the same rules as CS?,as in you have 1 life and once your dead thats it for the round, or do you just respawn and run back into the fray?

Biggles
16-03-2005, 11:50 PM
No - that's what made it better for a small group of players, even using bots. When you die you respawn back at the respwn location (which are made in the map so that the enemy cannot enter/camp them - at least if the mapper did it right). So you get back in almost instantly (short delay). Objective sare usually a series of flag points so the two teams are pushing for control of ground. The mapper can make it so that the flag needs moer than one player to capture, and control the delay before a cpature is complete. It's much better than CS, IMHO. None of that waiting around for minutes while the last AWP camper hides ina dark corner, and team play always wins the day.

Sam I Am
17-03-2005, 08:53 AM
Care to share any of the cs source stuff?

Im admin on a CS Source server and we are always looking for good maps.

Biggles
17-03-2005, 09:29 AM
Well if I ever get my current map finished, sure. It's my first Soruce maps so it's a bit of a learning experience.