About to press sumbit order...

[goobernoodles]

I'm about to upgrade to the following:

-Gigabyte GA-P67A-UD4-B3
http://www.newegg.com/Product/Produc...82E16813128478

-Intel Core i5-2500K
http://www.newegg.com/Product/Produc...82E16819115072

-CORSAIR Vengeance 8GB DDR3 1866Mhz
http://www.newegg.com/Product/Produc...82E16820233142

Anyone see any reason to change the RAM/Mobo?

 

[aphelion02]

RAM speed is irrelevant for sandy bridge, go for 1333 instead. Everything else is overkill. Also make sure you are not using a heatsink that is incompatible with the ram heatspreaders.

 

[goobernoodles]

RAM speed is irrelevant for sandy bridge, go for 1333 instead. Everything else is overkill. Also make sure you are not using a heatsink that is incompatible with the ram heatspreaders.

Is this because overclocking is multiplier based rather than FSB?

 

[aphelion02]

Yes, and tests show only 1-2% performance increase at most in any application that isn't specificially designed to test memory.

 

[Absolution75]

Indeed change your ram.

I've bought 16GB of PC1600 for the price of that ram. I think there is a newegg shellshocker today for 16GB for $130.

 

[goobernoodles]

I ended up going with the same Crucial line - just the 1600Mhz version for $90.

http://www.newegg.com/Product/Produc...82E16820233144

Can't wait!

 

[BTRY B 529th FA BN]

Nice

:thumbsup:

 

[BTRY B 529th FA BN]

...wait, do you have a Micro Center near by? The i5 2-500K chips are 179$ there

 

[Bartman39]

Yes, and tests show only 1-2% performance increase at most in any application that isn't specificially designed to test memory.

Which test(s)...?

Just ran Aida64 and quite a bit of difference in running 1333mhz Vs. 1866mhz...?

 

[fffblackmage]

Which test(s)...?

Just ran Aida64 and quite a bit of difference in running 1333mhz Vs. 1866mhz...?

How well does Aida64 reflect realistic scenarios?

The Best Memory for Sandy Bridge - bit-tech.net

In general, it's not enough for anyone to notice the difference. The best real-world scenario where the faster ram made a difference was the "mPlayer and 7-Zip Multitasking Test." Not really significant, IMO.

I ran my own tests with my current rig comparing DDR2-800 (CL5) and DDR2-1000 (CL5) and I got an average 1.75% FPS gain.
(Resident Evil 5, Lost Planet 2, STALKER:COP bench, Far Cry 2)

While faster ram is definitely preferred, it's just not going to make a significant difference. I may spring a few bucks for DDR3-1600 over DDR3-1333, but I certainly wouldn't go out of my way to get DDR3-1866, especially if the ram requires anything over 1.5V.

 

[Bartman39]

While I understand what your saying for most part... I think that the reason we OC and run faster than stock settings is to get the most out of our systems...? If you can run faster memory then why not...? If I had of known I would have got even faster memory like some 2133mhz DDR3... Also as for FPS differences I dont see that as being much of an indicator as its more based on the GPU and CPU more than memory speeds... Sure some programs will not have much benifit from the faster memory but for sure alot of them do... In the real world even with a fast CPU & SATA III SSD like whats in my sig Windows doesnt really load much faster than my I3 laptop which also has an SSD that is real old... But go do some video rendering and watch what happens, I view it as a high HP setup needs a high performance fuel system to work with it (meaning faster ram)... I might try some different programs video and various others and see if different memory speeds has much of an affect...? I dont expect to see a 30-40% difference but 5-10% is better than the average bear...

Also something to note... Speed is nothing more than a question of money...

BTW the memory in my sig is low voltage to begin with...

I did read through the bit-tech.net article and if you read the end conclusion for most 1600mhz is fine they say but 1866mhz is the shizz... I think though 1333mhz is a bit on the slow side IMPO... Good article BTW...

 

[aphelion02]

You could argue that for that money he should have gone for a 2600k, better mobo / gpu, or more RAM. There's so many places to upgrade before you consider upping RAM speed.

 

[mfenn]

Also something to note... Speed is nothing more than a question of money...

This is the key point. If you have infinite budget and want to build the fastest thing humanly possible, then knock yourself out with high-speed memory. The rest of us would rather spend that $60 upgrading from a 6950 to a 6970 (for example).

 

[Bartman39]

This is the key point. If you have infinite budget and want to build the fastest thing humanly possible, then knock yourself out with high-speed memory. The rest of us would rather spend that $60 upgrading from a 6950 to a 6970 (for example).

If I had an infinite budget then the build would have been alot different but with a bit of working you can have better by just planning for it which looks like what the OP is... While I understand your thinking of saving him a little money are you not also leaving some performance on the table by doing so...? Just to point out the money I saved buying 6950HD`s (2gig ref`s) and unlocking them to the 6970 spec did allow for the high-speed memory in my case (could also be his)... I too would have bought the Vengance ram if it had been in stock at the time but only the Gskill was with the voltage I wanted... I figure if you go run of the mill then you will for sure wonder later what could have been...

 

[aphelion02]

Once again, the OP would be much better served by going with more RAM, saving up for an SSD, or a set of faster GPUs before even considering RAM speed.

 

[mfenn]

Once again, the OP would be much better served by going with more RAM, saving up for an SSD, or a set of faster GPUs before even considering RAM speed.

:thumbsup: Exactly. Cost-benefit analysis of each upgrade is the best way to get a rig with the highest overall performance. Who cares if you really whoop ass in some synthetic?