| Intel core i7 3770k/i7 3770/i5 3570k for advanced calculus Hi,
Im buying a new PC for more difficult calculations in Matlab, Maple or R.
Only component Im not 100% sure is CPU. I will have 'ASUS P8Z77-M PRO' motherboard.
One guy told me to buy i5 3570k - after overclocking it will be better and cheaper than i7 3770. But i7 3770 has 8MB L3 cache and uses hyperthreading (i5 3570k has only 6MB L3 cache and doesnt use HT).
So I thought that i7 3770k would be the best choice, but after I read some reviews Im not sure because of high energy consumption and high temperatures.
What do you think? Thank you for all advices. |
|
 KrisnatharokCaveat EmptorPremium
join:2009-02-11
Earth Orbit
kudos:7 | Does your calc program take advantage of multiple CPU cores or only run on one? If multi-threaded, go for the i7-3770K. If not, get the i5-3570K and overclock it.
The power draw of both the i7-3770K and the i5-3570K is identical (77 W), one does not consume more power than the other, although the i7 may run a little hotter due to the HT (probably not enough for you to care, however, unless you are an extreme overclocker).
Whichever CPU you go for, grab an aftermarket cooler--the Intel one that comes with the CPU is crap. Something like the $20 Hyper 212+ is all you need.
--
If we lose this freedom of ours, history will record with the greatest astonishment, those who had the most to lose, did the least to prevent its happening. |
|
Reviews:
 ·Frontier Communi..
 ·WildBlue
| reply to TomCZek
When doing heavy math calculations you really want to watch your heat. Even though the program does not error out you may still get incorrect answers because of a heat stressed cpu or memory.
I would definitely run superpi several times to make absolutely sure you are getting accurate numbers even if you are not overclocking.
Heat stressed components are fine when playing games. If you are off by .001% you might miss your target or get shot etc. But you do not want to be off by even .0000001% if you are doing critical calculations. |
|
 DarkLogixTexan and ProudPremium
join:2008-10-23
Baytown, TX
kudos:3 | reply to Krisnatharok
IIRC matlab does make use of multi-core
though my only exposure to it was in working with co-workers using it and seeing that the dual socket 2GHz (ie 8 cores plus HT) was able to crunch the numbers faster than the single socket 2.6GHz (quad w/HT)
For calc you'd not want to OC as OC'ing would put heat stress on the CPU and make it more likely to give slightly wrong output.
I'd run ycruncher set to the largest amount your amount of ram can store, or put it in stress test mode. |
|
KrisnatharokCaveat EmptorPremium
join:2009-02-11
Earth Orbit
kudos:7 | said by DarkLogix:though my only exposure to it was in working with co-workers using it and seeing that the dual socket 2GHz (ie 8 cores plus HT) was able to crunch the numbers faster than the single socket 2.6GHz (quad w/HT)
Eight cores plus HT? On a single CPU? I think you meant six, as in a dual socket 1366 board?
--
If we lose this freedom of ours, history will record with the greatest astonishment, those who had the most to lose, did the least to prevent its happening. |
|
DarkLogixTexan and ProudPremium
join:2008-10-23
Baytown, TX
kudos:3 | I said dual socket.
4 cores w/HT per socket
2 sockets = 8cores plus HT (16cores) |
|
|
|
KrisnatharokCaveat EmptorPremium
join:2009-02-11
Earth Orbit
kudos:7 | So why go with two quad-cores when you can go with two-hexacores?  |
|
DarkLogixTexan and ProudPremium
join:2008-10-23
Baytown, TX
kudos:3 | It was what was on our HP contract at the time. |
|
 OctaveanPremium,MVM
join:2001-03-31
New York, NY
kudos:1 | reply to TomCZek
That takes me back. I remember running Maple in college at home and in the math lab. As well as my old Ti-82 graphing calculator. The system requirements back then weren’t very demand for Maple so I am guessing Matlab and R may need more processing power.
Is this for work or are you a student?
In the back of my mind I tend to think that a lot of people working in the field of math / science would simply buy a Mac and be done with it. |
|
