Intel i5-760 vs i5-2500k
- KVRAF
- 1706 posts since 22 Apr, 2009 from Belgrade
deciding on a processor for a DAW upgrade - i stopped following the computer market for a few years and i could definitely use some help here.
so, i've almost decided to pick an Intel CPU. the budget is under 200 EUR.
now, here's the thing - i can't choose between the i5-760 and i5-2500k. the 760 has more cache, but the 2500k is clocked higher (3.3 GHz vs 2.8 GHz) and can be overclocked.
which one is the better choice for a DAW (disregarding the price difference of 30 EUR).
also, any other DAW building tips are appreciated - if you can recommend any specific motherboard, RAM, or anything else, please do!
i'll be running Win XP SP3.
so, i've almost decided to pick an Intel CPU. the budget is under 200 EUR.
now, here's the thing - i can't choose between the i5-760 and i5-2500k. the 760 has more cache, but the 2500k is clocked higher (3.3 GHz vs 2.8 GHz) and can be overclocked.
which one is the better choice for a DAW (disregarding the price difference of 30 EUR).
also, any other DAW building tips are appreciated - if you can recommend any specific motherboard, RAM, or anything else, please do!
i'll be running Win XP SP3.
Bedroom Producers Blog << Free VST Plugins!
-
- KVRAF
- 1924 posts since 15 Oct, 2008 from Germany
I would probably go with the 2500k because of the newer architecture.
Here's a comparison of the two CPUs: http://www.anandtech.com/bench/Product/191?vs=288
Interestingly there's no clear winner in the benchmarks and I can't really say anything about the performance in a DAW system. Maybe others have more specific experience. But the 2500k is certainly a good CPU.
But you should probably be thinking about using Win 7 instead of XP to make best use of the hardware.
Here's a comparison of the two CPUs: http://www.anandtech.com/bench/Product/191?vs=288
Interestingly there's no clear winner in the benchmarks and I can't really say anything about the performance in a DAW system. Maybe others have more specific experience. But the 2500k is certainly a good CPU.
But you should probably be thinking about using Win 7 instead of XP to make best use of the hardware.
- KVRAF
- Topic Starter
- 1706 posts since 22 Apr, 2009 from Belgrade
thanks for the tips. i think i read somewhere that cache is more important for DAW use than the clock, but i'm really not sure anymore. guess i'll go for the 2500k, but i'll give it a few more days of thought before taking action.
damn, i used to know everything about the latest pc components until a few years ago, now i'm like a damn grandpa.
damn, i used to know everything about the latest pc components until a few years ago, now i'm like a damn grandpa.
Bedroom Producers Blog << Free VST Plugins!
-
Lunatic_Machinist Lunatic_Machinist https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=192752
- KVRist
- 297 posts since 2 Nov, 2008
my vote goes for the 2500k, newer architecture, faster clock, what's not to like?
also, take a look at the posted benchmarks. 2500k thrashes the older cpu in practically every category but one.
also, take a look at the posted benchmarks. 2500k thrashes the older cpu in practically every category but one.
- KVRAF
- 4141 posts since 11 Aug, 2006 from Texas
I really recommend against overclocking. It can (and often does) introduce strange stability problems to the system. These problems often increase in frequency the longer the system is overclocked too. The last thing you want is for your system to crash while you're in the zone and making the final tweaks to a lead line. The i5-2500k is a great CPU and I'd recommend that one. The price difference between the locked and unlocked variant is quite small (about $15 USD right now) that it's up to you to pick one over the other.
But really. Don't overclock. You'll be glad you did later.
But really. Don't overclock. You'll be glad you did later.
- KVRAF
- Topic Starter
- 1706 posts since 22 Apr, 2009 from Belgrade
good point!bmrzycki wrote:I really recommend against overclocking.
thanks for the tips everyone, i'm almost sure i'll go for a 2500 now. glad i asked.
Bedroom Producers Blog << Free VST Plugins!
-
- KVRist
- 434 posts since 22 Nov, 2010 from Argentina
Then you overclock wasnt stable and you need to tweak it further. Its always been like that. If you're not willing to tweak until perfection (thus, stability), then you shouldnt OC at all. It would be dangerous to do otherwise.bmrzycki wrote:I really recommend against overclocking. It can (and often does) introduce strange stability problems to the system.
Between those two, 2500K hands down. But, next Intel CPUs are really around the corner (two to three months), and it seems that they will directly replace all the i5s in performance and price segment, plus new mobos with (hopefully) new features. PCIe 3.0 is one (not that it matters much in DAWs).
So its your call, if you want something now, i5 2500K is the best choise.
I think you're looking at the bars without reading their descriptions. The 2500K wins each single time.paterpeter wrote:Interestingly there's no clear winner in the benchmarks
-
- KVRAF
- 1924 posts since 15 Oct, 2008 from Germany
Oops, you're right, I didn't read it correctly. I just though: larger=better.The Chubu wrote:bmrzycki wrote:I think you're looking at the bars without reading their descriptions. The 2500K wins each single time.paterpeter wrote:Interestingly there's no clear winner in the benchmarks
- KVRAF
- 4141 posts since 11 Aug, 2006 from Texas
It all depends on if you want to spend your time on. If you enjoy playing with bios settings, reading forums, and running benchmarks to try to get ~10-15% more out of a CPU then be my guest. I'd much rather have a system with all my parts stress-tested by the manufacturers for a specifically designed load.The Chubu wrote:Then you overclock wasnt stable and you need to tweak it further. Its always been like that. If you're not willing to tweak until perfection (thus, stability), then you shouldnt OC at all. It would be dangerous to do otherwise.bmrzycki wrote:I really recommend against overclocking. It can (and often does) introduce strange stability problems to the system.
And on the comment of stability, there are 2 kinds of problems overclocking can create:
1. Immediate problems - these are the kinds of problems that are easy to find and are often related to clocking beyond the speed at which the electrical signal stabilizes in the slowest logic block of a given chip. You just walk the clock down until things work.
2. Mid-term problems - when you overclock you are running the part in a way that either (A) exceeds QA testing yields or (B) was never designed to run in. This will degrade the life of the part. The big question-mark here is if that degradation is in the months or years. You might spend a sold week setting up the perfect stable overclocking scenario. Then 6 months from now you start getting strange crashes. Your stable setup is no longer stable.
That's why I said to not overclock. I am assuming the OP wants to use his computer to make music, not to spend all his time trying to get the last few percentage points of performance out of it. If you or others want to do that then more power to you. I personally see it as a lot of time invested for relatively low output. If I really need the extra performance I'll save my money and buy the better part.
-
- KVRist
- 434 posts since 22 Nov, 2010 from Argentina
So? "Can create" is not "will create" you know? I got my C2D over volted and over clocked since year and a half (replaced by a 2500K recently) and my previous GPU also.bmrzycki wrote:And on the comment of stability, there are 2 kinds of problems overclocking can create:
Tell me, how many CPUs have you dealt with that were burned after a year or two of OCing? Myself, none at all. Either it works, or it doesnt. Hell, just go to any hardware tech sites, for every user with a burned GPU or burned CPU thread, you have 10 answering to him/her that already have their hardware OCed and working perfectly.bmrzycki wrote:2. Mid-term problems - when you overclock you are running the part in a way that either (A) exceeds QA testing yields or (B) was never designed to run in. This will degrade the life of the part. The big question-mark here is if that degradation is in the months or years. You might spend a sold week setting up the perfect stable overclocking scenario. Then 6 months from now you start getting strange crashes. Your stable setup is no longer stable.
The big question mark is how much time do you think that a CPU can last wo OCing and how much time do you think OCing can take out of the CPU?
My guess is that you think that OCing is a death sentece for a CPU, a thing that has been done and done over the years since the first commercial CPU came out to the market. Do you think that ppl would still be OCing their CPUs if that was the case?
http://forums.guru3d.com/showthread.php ... ost4213664
Here, picked up random thread of Guru3D, only on that page you got like 10 users with OCed components, half of them with both oced CPUs and oced GPUs, some very new (3890X) some not so new (i5 750, i7 920) and one user with an old cpu (Q9xxx) at a pretty high OC.
Now go and watch ALL other hardware subforums, you wont find even 10 threads bout burned stuff cuz bad OC on their first pages. There is even a guy who ran furmark for 3hs cuz he didnt knew what it was an endless stressing test. Did his card explode or something? Not at all (i did have a good laugh though lol).
Not even Intel and AMD think that it can take too much life out the CPUs since they already OC their CPUs dinamically. Not counting that most iterations of a single generation of CPUs are based on the same chip. That all new and shiny 2700K isnt some revision of the Sandy Bridge CPU, its just a 2600K clocked higher.
And no, im not saying that you wont experience any problem. But a "you might experience issues" isnt a good reason for not OC either. Hell, if i minded so much the various "you might" s that there are around, i wouldnt even try to build a PC, after all, i might break it! lol
Where did the OP said that he has only a few hours to live? Maybe a day or two? lol Now you're just trolling.bmrzycki wrote:That's why I said to not overclock. I am assuming the OP wants to use his computer to make music, not to spend all his time trying to get the last few percentage points of performance out of it
- KVRAF
- 4141 posts since 11 Aug, 2006 from Texas
Sigh, I'm done.
I've built computers since 1993, work for a fabless semiconductor company as a software engineer and have overclocked (and underclocked) many cpus across different architectures. I've done QA and bringup on CPUs/boards. I've given my opinion on the matter and so have you.
Like I said, do what you want with your own equipment and free time.
To the OP, I'm sorry the thread has been derailed. I won't post further.
I've built computers since 1993, work for a fabless semiconductor company as a software engineer and have overclocked (and underclocked) many cpus across different architectures. I've done QA and bringup on CPUs/boards. I've given my opinion on the matter and so have you.
Like I said, do what you want with your own equipment and free time.
To the OP, I'm sorry the thread has been derailed. I won't post further.
-
- KVRian
- 1392 posts since 28 May, 2008 from Saint Paul, MN
Forget i5! You NEED i7 with hyper-threading
Seriously though, the price of current i7s should come down a bit this March (when new ones are released) so if you can hold out, I would go for the "Sandy Bridge" i7 3.4 Ghz ($299 US right now).
Something to consider...
R
Seriously though, the price of current i7s should come down a bit this March (when new ones are released) so if you can hold out, I would go for the "Sandy Bridge" i7 3.4 Ghz ($299 US right now).
Something to consider...
R
-
- KVRAF
- 1959 posts since 4 Nov, 2004 from Manchester
Bang per buck the 2600 is better but the 2500 is no slouch. I wouldn't go lower than a 2500 through.
Just to throw my 2 cents in on the overclocking, the current generation sees about a 30% increase in DAWBench performance on a modest overclock. When I say modest I mean as far as you can push it without raising the voltages from stock which is still in the mid 4Ghz. Modern CPU's are rated for 5 - 10 years at stock voltages so not overly concerned with it causing an issue as the is a lot of overhead in the silicon. We've been selling and supporting overclocked systems for the best part of a decade now and I can count on one hand the number of CPU's failed in that time frame out of the 10's of thousands we've shipped.
Just to throw my 2 cents in on the overclocking, the current generation sees about a 30% increase in DAWBench performance on a modest overclock. When I say modest I mean as far as you can push it without raising the voltages from stock which is still in the mid 4Ghz. Modern CPU's are rated for 5 - 10 years at stock voltages so not overly concerned with it causing an issue as the is a lot of overhead in the silicon. We've been selling and supporting overclocked systems for the best part of a decade now and I can count on one hand the number of CPU's failed in that time frame out of the 10's of thousands we've shipped.
-
- KVRAF
- 4056 posts since 8 Jan, 2005 from Hamilton, New Zealand
It does look like. Those are all solid points, and every machine I own has been overclocked from day 5. Cpus do degrade, but the length of time you own a given computer for will, for most "constant upgraders", be much shorter than the period of time till degradation.The Chubu wrote:Doesn't looks like.bmrzycki wrote:and have overclocked (and underclocked) many cpus across different architectures
Cheers,
Matt
I make music: progressive-acoustic | electronica/game-soundtrack work | progressive alt-metal
Win 10/11 Simplifier | Also, Specialized C++ containers
Win 10/11 Simplifier | Also, Specialized C++ containers
