i5 or X6 for programming (no gaming). Please suggest

[xpanta]

Hi,

I am going to upgrade soon. I now have a Dual Core E8500 + 4GB RAM on a Gigabyte EP45 mobo.

90% of pc usage is programming and web development. I am in between intel i5 and AMD X6 processor. Which one is better?

Thanks.

 

[Arachnotronic]

Hmm, tough. Which x6 and which i5?

 

[xpanta]

Hmm, tough. Which x6 and which i5?

don't care about the "subtype" of the cpu. The technology is what I am interested in. Do more cores (X6) give better programming experience, or not? (considering the price difference, of course)

 

[Arachnotronic]

Well, more cores is actually pretty good for compilation, and in this benchmark, the 1100T spanks the 2500K. http://www.overclock.net/hardware-news/1143169-amd-b3-bulldozer-stepping-listed-family.html

 

[Broheim]

honestly it really doesn't matter much, but the i5 will offer you the best performance (if we're talking SNB). singlethreaded performance > more cores during compilation(this depends on which language you're programming in and what compiler you're using).

 

[aigomorla]

X6 more cores... means more resources u have available.

If its what you ask, you dont need uber fast processing power, just massive threads.

The X6 will give u the most threads / $$$

honestly it really doesn't matter much, but the i5 will offer you the best performance (if we're talking SNB). singlethreaded performance > more cores during compilation(this depends on which language you're programming in and what compiler you're using).

Thats only if he wants to have his programing being the only thing that is open, and also on what OS.

But if the OS can handle a lot of cores... it means he can have more background programs running by having more working cores as long as the OS can handle the offload.

 

[DaveSimmons]

What tools?

Faster per-core per performance with i5 may make the UI more responsive in IDEs like Visual Studio, and the tools you use might or might not benefit from the extra cores in an X6.

 

[aigomorla]

guys realistically he's not going to see that big of a difference between an i5 and a x6 in single thread performance...

the X6 isnt Bulldozer.

 

[DaveSimmons]

^ true, I'm still using a 2.4 GHz q6600 at work, though it could be snappier in VS2010 so I'll be nagging my boss for an i5 upgrade soon.

Power use and noise with a stock fan favors the intel, but the X6 can be cooled quietly with an aftermarket HSF.

I guess the answer is: both are pretty good, do what you like.

 

[antisocialmunky]

If you're doing multi-core stuff, won't you have to optimize for HT ?

 

[Maximilian]

Maybe It is a very good chance to get Mount Winged Guardian in WOW!! My friend is crazy about WoW and he told me that he has taken part in 2011 Battle.net Invitational WOW 3V3 TOURNAMENT. He found there's an activity being held by ********* is interesting. If you guess correctly which teams will be the top three champions in 2011 Battle.net Invitational WOW 3V3 TOURNAMENT, you will get prizes! He is interested, because he's one of the ten teams and he may vote for his team! I also wanna join it and vote for these teams!

 

[DaveSimmons]

If you're doing multi-core stuff, won't you have to optimize for HT ?

In general Windows applications, a single process doesn't know or care whether its threads are hyperthreads or cores. Threads are used for things like maming the user interface more responsive by keeping its thread separate from some long task running in another thread. That kind of thread programming works even on single-core non-hyperthreaded CPUs.

Parallel programming for bulk data crunching is a different topic.

Making your development tools responsive is a third topic.

 

[antisocialmunky]

204 more posts before you find out if Anand uses signed 16 bit integers for post counts.

 

[mb103051]

x6 for what your gonna be using it for....best option is i7 970/980 6 core but amd x6 is a good deal ,best bang for the buck...

 

[gevorg]

Here is one compiling comparison, the 2600K looks like a champ for the price, the 2500K is a little behind the 1100T. A modest overclock can push the 2500K quite a bit higher.

http://www.anandtech.com/show/4083/...core-i7-2600k-i5-2500k-core-i3-2100-tested/19

 

[Voo]

There are extremely large differences depending on WHAT exactly you do. The conventional compilation process is rather easy to multithread and can for large projects take care of 6 cores or more.

Now if you're using LTCG and its variants in other compilers it looks quite different: The linking process is pretty much single threaded, which sadly also includes LTCG. MS is working on improving that (and afaik it's not much better in gcc), but at the moment it would greatly prefer high single threaded performance.

 

[Dadofamunky]

204 more posts before you find out if Anand uses signed 16 bit integers for post counts.

Antisocialmunky, behave yourself! Back to your room! ~Dadofamunky

 

[WaitingForNehalem]

204 more posts before you find out if Anand uses signed 16 bit integers for post counts.

Having it set to a signed 16 bit integer wouldn't make since post counts can't be negative .

 

[Matthiasa]

Not counting the fact certian people have over(or close to it) the unsigned 16 bit count...

On topic though its basically going to be as already stated by others.

 

[xpanta]

Here is one compiling comparison, the 2600K looks like a champ for the price, the 2500K is a little behind the 1100T. A modest overclock can push the 2500K quite a bit higher.

http://www.anandtech.com/show/4083/...core-i7-2600k-i5-2500k-core-i3-2100-tested/19

Thanks for tip, that was very insightful.

 

[pmccall2]

It depends on HOW you work too, everybody does things differently.

On almost any project Im usually running many things at once. On almost anything I can have many programs open (Eclipse, Photoshop, svn, IRC, a VM, plus many browser windows, and maybe winamp for some tunes =p). I dont compile too often, and frankly compile time has never been a big issue for me. what is important to me is how responsive the system is when im working.

I switched from a q6600 to a 1090t about a year ago, and im happy with it. If fast single-threaded compiles are a major issue you will probably need the intel part. In almost any other case (especially if you are running a vm) I would recommend the 1090t.

 

[MisterMac]

I do DLL projects with unmanaged coding, and hooking.

I'd say single threaded performance unless your have a HUGE ASS project, would be key first, seeing as it's not like there's a huge core difference from 4 SB to 6 Thubans.
(Especially with LTCG)

BD Would most likely fail further.

2600k seems the winner, unless you can wait AND afford the Hexacore SB-E's.

They seem to indicate some more small improvements over the D0 steppings of 2500/2600?

If you do have huge projects most likely youll divide into subprojects within whole solution anyway, and a full compile equals coffee time no?


I personally don't think we're at the total more cores more performance overall yet in direct compiling with whatever IDE.
Give it some years tho.....

2600k + a not so expensive Aftermarket cooler give it 4,5 ghz clock and go

 

[Arkaign]

I like the X6s and the 2500/2600k quite a bit. Sounds like you have a choice between winning and winning. The % difference in those kinds of workloads is gonna be minor in either direction, so maybe just get the best setup for the $$?