Q6600, what memory for midrange OC'ing?

B

Buztafen

Limp Gawd
Joined
May 19, 2005
Messages
238
Afternoon all,

Im upgrading my current sig computer with the following:-

Q6600 CPU
Gigabyte GA-EP45-DQ6 Mobo http://www.gigabyte.com.tw/Products/Motherboard/Products_Overview.aspx?ProductID=2831
4870 GPU.

Ive got a Zalman CNPS 9500 which should fit nicely on the mobo/cpu http://www.zalman.co.kr/eng/product/Product_Read.asp?idx=160...
And a 600w Tagan PSU http://www.eclipsecomputers.com/product.aspx?code=PST-TG600U33

What memory should i go for to get some decent 'middle of the road' OC'ing from the Q6600?:confused:
Im thinking of getting 4gb as a minimum and will be running XP and possibly Vista.

Cheers.
 
All you need is PC6400 to get a Q6600 to 3.6ghz. 9x 400 FSB at a 1:1 is DDR800.
 
Anything DDR2 800 - but whatever you get, make sure it runs as close to 1.8V (JDEC standard) as possible. That way if you want to overclock above 400Mhz you'll probably get it with a bump in voltage. A lot of the kits out there run at 2.0 or 2.1 by default so they don't give you much overhead (2.1V is the max).
 
SharpieFiend said:
Anything DDR2 800 - but whatever you get, make sure it runs as close to 1.8V (JDEC standard) as possible. That way if you want to overclock above 400Mhz you'll probably get it with a bump in voltage. A lot of the kits out there run at 2.0 or 2.1 by default so they don't give you much overhead (2.1V is the max).
Click to expand...

Thanks for the heads up on the voltage. Think im going for something like these:-

http://www.b4udirect.com/index.cfm?course=/home/product/00566069/&bhcp=1

Theyre on a 1.9v...or maybe these:-

http://www.b4udirect.com/home/product/00149437/.htm

or these....

http://www.overclockers.co.uk/showproduct.php?prodid=MY-019-GS&utm_source=froogle

in fact the above set are at a good price...anyone see any problems with them?
 
Well ive just gone for the bottom set......whats so bad about them? Is the world gonna end when i put them in my mobo? Will they call me a b*tch behind my back?:confused:
 
Just interested to find out whats wrong with them as ive heard that theres not much real world difference between CL4 and CL5 memory if im just playing games and HD content...

Not being funny or owt ;)
 
Buztafen said:
Just interested to find out whats wrong with them as ive heard that theres not much real world difference between CL4 and CL5 memory if im just playing games and HD content...

Not being funny or owt ;)
Click to expand...

There wouldn't be a huge difference, but there'd be a difference for sure. In any case, unless you're a hardcore gamer, you probably won't care.

EDIT: Hell, you may even be able to get those timings down to 4-4-4-12 anyway. I'd at least try.
 
Well i'll see what they can do first then if im not happy i'll sell and upgrade....though the seeds of doubt have been sown, damn you jimbo!!!! ;)
 
Buztafen said:
Well i'll see what they can do first then if im not happy i'll sell and upgrade....though the seeds of doubt have been sown, damn you jimbo!!!! ;)
Click to expand...

Nah, don't listen to me. I'm sure you will be satisfied. Once you get them, just toss them in the PC and fire up your favorite game. Then, if you want, see if you can run them at 4-4-4-12. I think at DDR2-800, you might be able to.
 
jimbo6846 said:
Nah, don't listen to me. I'm sure you will be satisfied. Once you get them, just toss them in the PC and fire up your favorite game. Then, if you want, see if you can run them at 4-4-4-12. I think at DDR2-800, you might be able to.
Click to expand...

Cheers jimbo, i'll give it a go....its only £50 at the end of the day.

Seen as ive got your attention heres another question:-

Ive got both XP and Vista licences available to me....which would be the prefered install?

Is Vista working 'properly' now?:rolleyes:
 
I have Vista x64 installed on my PC. So far it's been working just fine for me. Of course, I don't do much but the occasional gaming and web surfing. There's really no reason to go back to XP, but at the same time I'm not really seeing a reason to switch to Vista. If you're going to be doing a fresh install on your system, I'd say go for Vista. You won't run into very many compatibility issues, if any at all. Just make sure you download the Vista versions (32 or 64-bit accordingly) of any programs you use, if you do. Other than that, I haven't seen any problems. Vista has definitely come a long way from the horror stories you have probably heard, much like all Windows OS's.

Also, be forewarned: If you upgrade the OS from 32-bit to 64-bit, you will need to do a fresh install. So back up any files you want to save if that's the case.
 
Ok cheers for the info, think im gonna go for Vista, can always change back if something important doesnt work.
 
Yeah, that was my reasoning too. Fortunately, no reason for me to change back, yet anyway.
 
How big a difference in performance is there really between 4-4-4 and 5-5-5? Does anyone have any links to benchmarks?
 
^^^See what youve started jimbo! People are panicking....this could be the end! :p

/waits patiently for his Q6600 and crappy 5-5-5-15 memory to arrive sometime today...
 
difference between 4-4-4 timings and 5-5-5 timings is measurable only in memory benchmarks. Not in real world performance.
 
Archaea said:
difference between 4-4-4 timings and 5-5-5 timings is measurable only in memory benchmarks. Not in real world performance.
Click to expand...

I never agreed with this. For some reason, some people like to say "real world performance" when what they really mean is they can't see any difference with just their eyes alone.
 
compare benchmarks on things that matter, like frame rates in games, application benchmarks etc, and there is no difference at all <1% at best. Run a synthetic RAM benchmark that only tests that one thing and and sure -- it'll show 20% faster.
 
if you are using programs that use huge amounts of RAM, you will notice a difference... I've got a couple I programmed that you can see the difference with small timing changes in RAM.

Or if you like to try something commercial.. use Mathmatica to due large calculations... that will show a difference as well.

For you normal, everyday programs.. you will not see a difference.
 
jimbo6846 said:
Some of the results on the other pages were much more favorable to tighter timings, too. Yes, some games don't benefit that much from tighter timings, but some do, and other applications/PC usage benefit from them as well, besides gaming.
Click to expand...

I think we are getting lost in the weeds here to the OP, but...I wouldn't say the typical ~ 1% is "much" Perhaps very little or nill is the proper thought train. I will ceede, having been around computers and building PCs since the mid 90's, that in days of old, CAS latency and timings made a much larger difference than they do today. When I researched this about a year ago before buying my lastest power rigg (Q6600 based) and saw multiple benchmarks showing zero difference or such a minor difference as to be just generic deviations/variations in test output I was surprised as well. For the Intel Quad and Dual cores, for the most part, CAS latency and timings just don't matter enough to care about.

If the RAM is the same price sure, buy the tigher timings. Typically it is not, and in that case don't lose any sleep over any of the RAM latencies in your preformance systems. The processor on that list that is most comparable to the Q6600 is the E6600. It's the same proc just two cores rather than 4 cores. Use that for your comparison, the results should be nearly identical to what you'll exp.
 
Archaea said:
I think we are getting lost in the weeds here to the OP, but...I wouldn't say the typical ~ 1% is "much" Perhaps very little or nill is the proper thought train. I will ceede, having been around computers and building PCs since the mid 90's, that in days of old, CAS latency and timings made a much larger difference than they do today. When I researched this about a year ago before buying my lastest power rigg (Q6600 based) and saw multiple benchmarks showing zero difference or such a minor difference as to be just generic deviations/variations in test output I was surprised as well. For the Intel Quad and Dual cores, for the most part, CAS latency and timings just don't matter enough to care about.

If the RAM is the same price sure, buy the tigher timings. Typically it is not, and in that case don't lose any sleep over any of the RAM latencies in your preformance systems. The processor on that list that is most comparable to the Q6600 is the E6600. It's the same proc just two cores rather than 4 cores. Use that for your comparison, the results should be nearly identical to what you'll exp.
Click to expand...

That's basically the same thing I've been saying. There's a difference, but you wouldn't notice it nor would most people care.

Personally, I prioritize higher clock speed rather than tighter timings. I may only get a slightly bigger boost than tighter timings grants, but that's fine with me.
 
jimbo6846 said:
That's basically the same thing I've been saying.
Click to expand...
Not exactly. I re-read your posts to ensure I wasn't setting up a straw man. I didn't want the Original Poster to be given slightly off information, thus my initial post...You, in fact said 5-5-5-18 timings were HORRIBLE. That's not true. 1% or 2% difference does not equate to HORRIBLE, it pretty much amounts to indifference, and might be partly attributable to testing variation. Lots o those tests reported are like 800 seconds total, and they shave off 11 having faster timings --- After 13 minutes do you think you'd notice 11 seconds? To your credit you aren't the only one in the thread, nor this forum who should see correction here. Also to your credit 5 to 10 years ago CAS latentency and timings mattered. But it's a myth that they matter today.

Personally, I prioritize higher clock speed rather than tighter timings.
Click to expand...

Here we go again...You can prefer and that's fine. I'd prefer faster timings and faster latencies as well, but only if the cost is the same, cause it just doesn't matter in the core 2 and quad chipsets like you'd think it would.

http://benchmarkreviews.com/index.p...ask=view&id=114&Itemid=1&limit=1&limitstart=5

Benchmark Conclusion: "In the end a 567MHz increase in system memory speed gave the framerate in World in Conflict only a 2 FPS improvement in the average framerate. While the minimum frames per second jumped from 13 to 18 for a 5 FPS improvement, the maximum was only affected by 1 FPS. This mediocre improvement is evidence of how insignificant the system memory speed is in relation to video game performance."

or this benchmark review --- comparing DDR3 to DDR2 --- it just doesn't matter. Arguably MHZ rating on the Core 2 Duo's and Quad Cores matter less than the timings! --- considering cost to performance ratios.

http://www.breakitdownblog.com/ddr2-800-vs-ddr3-1333-does-speed-matter/

video_encoding_ddr2_vs_ddr3.png


Normalized FPS taken over several games.

gaming_ddr2_vs_ddr3.png


Benchmark Conclusion:So the good news from all of this is that you can safely buy DDR2-800 ram for $40/GB instead of the f#*$(@$$ crazy DDR3 ram for the forseeable future. And don’t even bother with DDR2-1066 or higher DDR2 ram… I found benchmarks on those as well comparing to DDR2-800 that again showed almost no difference in performance.



Just today I had to tell someone I was wrong. I thought E85 was mostly gas and only partly ethanol. It's the reverse, it's only 15% gas. I admited I was wrong after being willing to wager most anything I was right. It happens.
 
The best testing I ever personally did was when 975 chipsets first came out. DDR800 at CAS5 vs DDR1066 at CAS5 was at best a 3%-5% increase in memory intensive apps. It had virtually no effect on games which is no surprise.

Latencies will become important again on i7 due to the integrated MCH but, they are not worth fretting over on Core CPUs and related chipsets.
 
Archaea said:
Not exactly. I re-read your posts to ensure I wasn't setting up a straw man. I didn't want the Original Poster to be given slightly off information, thus my initial post...You, in fact said 5-5-5-18 timings were HORRIBLE. That's not true. 1% or 2% difference does not equate to HORRIBLE, it pretty much amounts to indifference, and might be partly attributable to testing variation. Lots o those tests reported are like 800 seconds total, and they shave off 11 having faster timings --- After 13 minutes do you think you'd notice 11 seconds? To your credit you aren't the only one in the thread, nor this forum who should see correction here. Also to your credit 5 to 10 years ago CAS latentency and timings mattered. But it's a myth that they matter today.



Here we go again...You can prefer and that's fine. I'd prefer faster timings and faster latencies as well, but only if the cost is the same, cause it just doesn't matter in the core 2 and quad chipsets like you'd think it would.

http://benchmarkreviews.com/index.p...ask=view&id=114&Itemid=1&limit=1&limitstart=5

Benchmark Conclusion: "In the end a 567MHz increase in system memory speed gave the framerate in World in Conflict only a 2 FPS improvement in the average framerate. While the minimum frames per second jumped from 13 to 18 for a 5 FPS improvement, the maximum was only affected by 1 FPS. This mediocre improvement is evidence of how insignificant the system memory speed is in relation to video game performance."

or this benchmark review --- comparing DDR3 to DDR2 --- it just doesn't matter. Arguably MHZ rating on the Core 2 Duo's and Quad Cores matter less than the timings! --- considering cost to performance ratios.

http://www.breakitdownblog.com/ddr2-800-vs-ddr3-1333-does-speed-matter/

video_encoding_ddr2_vs_ddr3.png


Normalized FPS taken over several games.

gaming_ddr2_vs_ddr3.png


Benchmark Conclusion:So the good news from all of this is that you can safely buy DDR2-800 ram for $40/GB instead of the f#*$(@$$ crazy DDR3 ram for the forseeable future. And don’t even bother with DDR2-1066 or higher DDR2 ram… I found benchmarks on those as well comparing to DDR2-800 that again showed almost no difference in performance.



Just today I had to tell someone I was wrong. I thought E85 was mostly gas and only partly ethanol. It's the reverse, it's only 15% gas. I admited I was wrong after being willing to wager most anything I was right. It happens.
Click to expand...

Thank you for the clarification, but now you have me confused. Why would you say tighter timings don't make any difference, but then say you prefer tighter timings to higher clock speeds? There's no noticeable difference between them, other then in benchmarking, but yet you felt the need to ridicule my choice?

Why exactly do you prefer tighter timings if they make no difference?
 
I'm quoting my exact words from my post above.

"I'd prefer faster timings and faster latencies as well, but only if the cost is the same"


My point here has been to say timings didn't matter in real world. For all practical purposes as proven in benchmarks they don't...

You then, perhaps, changed your tuen to the fact that you prioritze Mhz over CAS and latency/(tighter timings)

I provided benchmarks that show that Mhz doesn't matter either in typical real world scenarios.


In answer to your valid point --- I am a partipant in [H]ardforum and by that token if I can get something that is 1/2 or 1% faster for the same price then sure. But I certaintly can't say 5-5-5-18 is horrible in comparison to 4-4-4-12 on the Intel plateform. Or Mhz is also important, when it really isn't. Perhaps steering someone to buy RAM that costs 3x's as much.

Clearly --- at anything other than basically the same price - the difference in cost for the performance gained isn't warranted.
 
Archaea said:
I'm quoting my exact words from my post above.

"I'd prefer faster timings and faster latencies as well, but only if the cost is the same"


My point here has been to say timings didn't matter in real world. For all practical purposes as proven in benchmarks they don't...

You then, perhaps, changed your tuen to the fact that you prioritze Mhz over CAS and latency/(tighter timings)

I provided benchmarks that show that Mhz doesn't matter either in typical real world scenarios.


In answer to your valid point --- I am a partipant in [H]ardforum and by that token if I can get something that is 1/2 or 1% faster for the same price then sure. But I certaintly can't say 5-5-5-18 is horrible in comparison to 4-4-4-12 on the Intel plateform. Or Mhz is also important, when it really isn't. Perhaps steering someone to buy RAM that costs 3x's as much.

Clearly --- at anything other than basically the same price - the difference in cost for the performance gained isn't warranted.
Click to expand...

Again, thank you for the clarification. A couple of points for the record:

1. I didn't "change my tune". I don't ever change my tune. I've always preferred speed over timings, because I have an Asus P5Q Deluxe board and Intel chipsets benefit more from looser timings/higher clockspeeds as compared to tighter timings/lower clockspeeds, as far as benchmarks are concerned.

2. I'd never tell anyone to buy overpriced RAM.

3. Mhz is important when OC'ing, that goes without saying. OC'ing RAM may not give much performance improvement, if any at all, but you can only take your system as far as your RAM will go. G-Skill RAM with loose default timings generally doesn't OC very well. The OP may not need to go over over DDR2-800, but then he again he may decide later that he does.
 
You must log in or register to reply here.
Back
Top