• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Best Processor for Compiling

G

Guy.Named.Matt

Limp Gawd
Joined
Jul 23, 2002
Messages
275
I have never really heard of any benchmarks comparing AMD and Intel processors' abilities when it comes to raw power and speed compiling code. I'm a linux junkie and compile almost everything from scratch (Gentoo Linux 2004). I also do quite a bit of programming and will be doing even more when I leave for University.

Anyways, I'm looking to build a system with the primary purpose of comiling code. My first instinct is to turn to the P4's with HyperThreading, but I thought I'd ask to see if anyone had any other opinions... not to start a flame war or anything. Thanks in advance.
 
read a few more cpu reviews. there are a few sites that always throw a compile or two in when testing cpus. none specifically come to mind, since im not a linux guy, but they are definetly out there... seem quite regularly.
 
3.4 EE is probably a demon at compiling, if you've got mad cash.

A cheaper way to get a significant speed boost would be to compile on 15 KRPM SCSI.

Get:

$99 - Adaptec 29160N - http://www.hypermicro.com/product.asp?pf_id=CTAD112&dept_id=13-003
$25 - Three-position twisted-to-flat SCSI cable - http://www.hypermicro.com/product.asp?pf_id=CAHM204&dept_id=01-002

and one of:

$161 - Fujitsu MAS3184NP 18.4 GB - http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=100097
$234 - Fujitsu MAS3367NP 36.7 GB - http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=100099
$491 - Fujitsu MAS3735NP 73.5 GB - http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=100101
 
The Pentium 4's strong points are more in the application areas like encodeing, rendering, multitasking, and in your instance compiling code.

The A64's strong points are more in gaming and quick access applications where timings come into affect. Its on-die controller is the main source of its performance. The Pentium 4 has its controller in the northbridge so in areas where timings are very important like gaming the A64 will have a slight edge from the low latency controller.

Depending on whether you are overclocking or not i would get a 3.0c or 3.2c on an Intel board if your not overclocking and an enthusiast board from Gigabyte, Abit, Asus, MSI if you are. 3.6Ghz is usually a pretty easy overclock with a high end Pentium 4.
 
Originally posted by STL
Ha! The Athlon 64 2.2GHz is $404 at NewEgg while the Northy 3.2C is $285 at NewEgg.

Remind me again why people buy AMD?
Click to expand...

Dude SHHH! lol

Lets not get anything started again.

The A64 is a great gaming processor and what i would pick if gaming was primarily all i did, although i still dont like their motherboards and shitty chipsets. I'm sure they will get better as time goes on.
 
i got a 3.0E [prescott] for $230 and now its @ 3.6 on crappy air.

running 56 idle and 70 load:eek: i know but i have watwe in the way
 
I recommend the Pentium 4 with HyperThreading. If you're doing a long compile you might find it handy to be able to go off and do something else, like read a webpage. Not that the AMD's can't do this, mind you, they just can't multitask as well.
 
Ok, thanks for all the advice. Looks like I'll be going with a Northwood after I price shop a little bit (I don't want to put out for the EE). I'm still a little while away from buying, however. As for the suggestion of boosting compiling speeds with a SCSI, in my case Hard Drive speed should not be the bottleneck, especially since I will be running SATA RAID. Thanks Again.
 
SCSI offers lower disk access times. RAID-0, god forbid, offers only higher STR. Learn what the difference is.
 
Originally posted by STL
SCSI offers lower disk access times. RAID-0, god forbid, offers only higher STR. Learn what the difference is.
Click to expand...

There are a number of issues involved. First, since STR is derived directly from the media transfer rate, its value also depends on what part of the disk is being read; larger outer cylinders have the highest STR, smaller inner cylinders have the lowest. Second, there's the matter of whether the access is really sequential.
Click to expand...

reference

guy asks about a processor, gets bs about drives in answer. I wonder how many of those 4k posts fall along these lines of useless.

I was bored so here is some things I dug up for you:

"There is a very real performance advantage in this benchmark to enabling SMP on a HT cpu. However, in the best case it only amounts to 11%. This means that if a specialised HT scheduler patch gained say 10% it would only amount to 1% overall - hardly an exciting amount. 1% should have been on the edge of statistical significance, but I haven't even been able to show any difference at all. This does _not_ mean there aren't performance benefits elsewhere, but they obviously need evidence."
Click to expand...

reference

I could not find anything solid on K8 gentoo compiling for AMD/Intel comparisons. I am sure they are out there but I have run out of patience.
 
Originally posted by Guy.Named.Matt
Hey, hey... the last thing we need is another flame war...
Click to expand...

Yea, he made the mistake of comparing the 3.2Ghz to the 3400+. The 3.4Ghz Pentium 4 is around the same price as the 3400+ so they aren't cheaper if you look at it that way. AMD usually follows suit to Intel's prices and lowers theirs slightly a few bucks because otherwise they wouldn't sell if they were quite a bit more then a Pentium 4 whether they are new tech or not just because of the fact most people buy Intel processors. AMD is more of an enthusiast processor that catches the eye of alot of people when they offer it cheaper then an Intel equivalent.
 
Compile time also depends on what you're compiling. I have a 40,000 line C++ program I developed that takes about a minute to compile on my system (see sig), and yet the C# project I'm working on that's currently 25,000 lines of code takes between 2 and 10 seconds to compile. I also noticed that large Java projects don't take very long to compile using NetBeans. I don't much care for Java though, I've just need it for a few classes, so don't take that as an endorsement for NetBeans (or Java).

So the choice between AMD or Intel may be moot if you're doing development in something other than C or C++, languages that have egregiously long compile times due to their large reliance on the preprocessor to manage external modules. Windows applications, in particular, take forever to compile because simply doing a
Code: 
#include <windows.h>
adds about a billion lines of code for the compiler to sift through. Why do you think Microsoft went to such extra trouble to implement precompiled header files? :) I compiled DScaler with and without PCH's, and the difference was amazing. Without PCH it took maybe 8 minutes, but with PCH it took less than one.

In either case I still recommend the P4 based soley on the presence of SMT (HyperThreading). Or get a dual workstation :D Either a Xeon or Opteron will serve you well in that regard ... if you can afford it.
 
[STL]
> Ha! The Athlon 64 2.2GHz is $404 at NewEgg while the Northy 3.2C is
> $285 at NewEgg.

[Benny Blanco]
> Go check the price of a A64 3200. Then go check the price of a
> 3.4ghz P4. Then come back and make another useless comment.

The 3.4C would nuke that particular Athlon 64.

I was comparing the Athlon 64 2.2GHz to the 3.2C because the latter outperforms the former in the aforementioned Linux kernel test, yet is less expensive.

Fool.

[M1ster_R0gers]
> guy asks about a processor, gets bs about drives in answer.

Compilation touches the drive.

Fool.

[burningrave101]
> Yea, he made the mistake of comparing the 3.2Ghz to the 3400+.

Because the 3.2C outperformed the "3400+" in the test given above.

I did not make the mistake...

> The 3.4Ghz Pentium 4 is around the same price as the 3400+ so they
> aren't cheaper

of using AMD's fake numbers.

See how you have been misled? AMD calls their procs by artificially high fake numbers so that you end up comparing them against the prices of /stronger/ and therefore more expensive Intel procs.

> AMD is more of an enthusiast processor that catches the eye of alot
> of people when they offer it cheaper then an Intel equivalent.

AMD misleads people.
 
You definitely met my expectation of useless.

You can pick and choose your tests and show the better chip in one bench, and say it owns all, but that only proves you are narrow minded, and you are misleading (or attempting to mislead) everyone who has the unfortunate experience of actually reading your post.

By the way, the game works both ways... (A64 3000 vs. P4 3.4EE).
 
Originally posted by STL
[STL]I was comparing the Athlon 64 2.2GHz to the 3.2C because the latter outperforms the former in the aforementioned Linux kernel test, yet is less expensive.
Click to expand...

Great, now as long as all we do all day is compile linux kernels, we know exactly which processor we should all buy! Excellent.

Erik
 
Lets try to keep this conversation to a higher maturity level guys and not start another flame war. We've had plenty of Intel vs AMD threads lately. Only state what you can prove with links. If you can't prove it dont bother typing a reply.
 
Originally posted by STL

Compilation touches the drive.

Fool.
Click to expand...

If you've ever compiled large projects then you'd know that while the drive is used, it is generally not the bottleneck. Open up a project in Visual C++ that is tens of thousands of lines of code and do a full rebuild. You'll see very high CPU usage (100% ... or 50% for a HT system, or 25% for a SMP+HT system like mine :)), but the hard drive light is hardly on all the time, or even the majority of the time. A lot of that CPU time is spent in the OS' cache manager, by the way, as a lot of the data from the disk is read over and over again due to C/C++'s old-fashioned preprocessor-dependent structure. This is especially true of compiling Windows applications where, as previously mentioned, doing a "#include <windows.h>" adds an inordinate number of lines of code for every source file that is compiled.

So yes, SCSI and 10k drives and whatnot would show gains in performance. But only a few percent in most cases, and not nearly enough to give even a spec of financial justification.

And lastly, I hardly see any reason to resort to ad-hominem attacks. Lighten up, STL. You obviously know a lot, and have an above average IQ around here, but that's no excuse to resort to such comments.
 
You must log in or register to reply here.
Back
Top