Well, assuming the Q9450 comes out on time, I'm getting that baby, and 4 gigs of DDR3 RAM (shoot me for teh overpricetyness)...
I'm on a Pentium 4 right now compiling at speeds of upto 30 seconds T_T
I'm under the assumption with the new quality of the Q9450 (assuming they don't screw it up like the phenom's), that I can compile in under 3 seconds, correct me if I'm wrong...I'm just sick of these long compile times...
A side note, I'm not just building a computer for the sake of shorter compile times, I need it for multiple reasons, this just happens to be an added bonus...
ID:182704
 Feb 28 2008, 9:10 pm
|
|
|
Feb 29 2008, 7:04 am
|
|
Nadrew
  |
Compiling what exactly? All languages use different compilation methods and that's really what determines how effective your system will compile stuff.
|
If your P4 is 2ghz, and your Quad core is 2.5ghz, then it should be roughly 25% faster, so your 30 seconds compile time should be ~20 seconds.
| |
It'll be BYOND, and I highly doubt an increase of 25%, intel's main deal isn't clock speed anymore...keep in mind, 4 cores, and single core on the Q9450 at 2.5ghz will be roughly 40%+ better clock for clock than a single P4 at the same speed...
Also, let's keep in mind, my P4 is at 3.0ghz right now...and I'll be overclocking my Q9450 to atleast 3.6ghz atleast. clock speeds aren't the only thing to keep in mind, they have L2 caches aswell, for which the new quads will have 12mb of, a huge amount, plus architecture... Since BYOND probably isn't multithreaded, I'd say within 5 seconds of compilation...though I could force all my cores on it for ubernessticness... | |
Moleboy wrote:
I'm on a Pentium 4 right now compiling at speeds of upto 30 seconds T_T 30 seconds is a long time? | |
You can't use multiple cores for a single core application, you will get whatever the 1-core vs 1-core performance boost is, but don't expect to go from 30 seconds to 4 seconds.
| |
Thank you for pointing out that common misconception.
I hear people talk about multi-core processors being so much better, and just want to slap them. I've got a dual core Pentium D 3.2ghz that I've overclocked to just south of 3.6ghz. Two of my buddies have quad core processors, and claim their PCs are better than mine. (both have 2.4ghz processors). I laughed in their faces when we did some benchmarking, as my two year old PC came out on top of almost every test. Upgrading just isn't worth it when something is brand new. You spend about three times what the thing is worth in the first few months, and it's just better to buy something that's about a year old anyway, as by then, the drivers stabilize almost any problem that has occurred with the hardware. Plus, there are enough reviews out there to point out if it is a good buy or not by then. My advice: during compile times, suck up the 30 seconds. It's a negligible amount of time. | |
um im running a windows vista home premium and i compile in about 3 seconds??? maybe ur computer just doesnt like u (my comp only cost me like $600 cnd and it was just with what i needed for schoolwork:P). so i guess buying expensive computers just isnt the right thing to do. also im using wireless net from my next door neighbour if that has anything to do wit it
| |
Why would compiling on your computer, with physical components to do the processing, be at all related to your neighbor's wifi?
| |
Strawgate wrote:
It depends on the project size you may be compiling Hello World, he may be compiling Byond The biggest factor has to do with how many maps you have. Compile time can go from 10 seconds to 10 minutes if you have as many maps as projects i deal with have. For test purposes to catch compile-time errors i simply uncheck maps. In the future itd be nice if maps didnt get recompiled unless they had been changed. However im going to guess this is harder than it sounds on the technical side. | |
I rarely put maps in my projects anymore. I'm a really big fan of using a modified version of Lummox's SwapMaps.
^_^ | |
Yes, I am aware of how multi-threaded applications work, I have a friend who is in fact learning to program them :-) but, I can force all cores on a program can't I?
Technically you can set a core to work/not work on something, and if I forced all cores to work on it, it wouldn't be like a multi-threaded application, but it might be a nice boost. And yes, I AM aware that you don't multiply the speed times amount of cores... Also though with the huge L2 cache, that'll definitely helpz. *slaps back* :D | |
Ter13 wrote:
Thank you for pointing out that common misconception. like I said, I'm getting the computer parts regardless, I just need to here about how long you think it will take me | |
Forcing multiple cores to work on a single thread will do absolutely nothing, as it is a single thread.
| |
This is another classic case of someone acting like they know what they are talking about.
The problem with this sort of conversation, is if someone does it in real life, they look like a moron after you point out that they don't have any clue what they are talking about, and either give up the debate, or further make themselves look like an idiot, get embarrassed, and decide it better not to talk for a while. Unfortunately, on the internet, the longer a thread carries on, the longer the person continues to back up their infactual/misinformed statement with more infactual/misinformed statements until either Godwin's law comes into play, someone locks the thread, or all intelligent parties decide it's not worth it, and the originating party believes they have won the argument. ^_^ I love teh intartubes! | |
Moleboy wrote:
but it might be a nice boost No. It will be absolutely 0% boost. There are several solid reasons for getting Dual/Quad core machines. Getting extra oomph out of single-core applications is not one of them. However, if you want to compile a large game, while browsing the internet, while playing World of Warcraft - Then you're going to get more smoothe behavior with a Quad core than a single core. The automatic load-balancing of the new Extreme quad cores is also nice (Automatically overclocks the used core(s) when not using all 4 cores), and the 12MB cache is a very noticable difference. | |
Did you know, with Hitlers war machine, the first computers were actually used to encrypt secret messages which took modern computers many years to be able to mimic?
Ter13 wrote: This is another classic case of someone acting like they know what they are talking about. | |
Strawgate wrote:
Did you know, with Hitlers war machine, the first computers were actually used to encrypt secret messages which took modern computers many years to be able to mimic? I knew that somone would do that once Godwin's law was brought up. George Gough | |
Moleboy wrote:
Yes, I am aware of how multi-threaded applications work, I have a friend who is in fact learning to program them :-) but, I can force all cores on a program can't I? See below for why you are not Technically you can set a core to work/not work on something, and if I forced all cores to work on it, it wouldn't be like a multi-threaded application, but it might be a nice boost. And yes, I AM aware that you don't multiply the speed times amount of cores... It would not as there is only one thread. I even see the application possibly screwing up. A single thread can be run on only one core at a time. You would probably set the cores to flop them around each-other (I guessing on that) and that is all. Also though with the huge L2 cache, that'll definitely helpz. That is for frequently used stuff, not really a big thing for actual running performance, just load IIRC. *slaps back* :D Riiiiiiiiiiiiiiiiiight George Gough | |