Which Ram speed is Best? Look here for the Benchmarks!!

[magnus2882]

Finally I can put this one to bed.

Some enthusiasts claim that because they can't run faster ram at the rated speed, that the same ram will actually perform better when downclocked - and at tighter timings. While i'm sure maybe SOME rams might behave this way, most simply will perform best when clocked at a higher frequency.

I've seen some benchmark's pulled off of Tom's hardware claiming that ram downclocked to 1600Mhz out performed the same modules clocked at 1800 and 2000. Let me just say I don't take much stock in benchmarks at tom's hardware- nor would I rely on them as gospel. I decided to run my own benchmarks, my own controlled environment, and use different software to confirm my results.

The Test Conditions:

- All benchmark's were performed right after a warm boot up.
- Each test was performed 3 times, and the best result was recorded.
- Bios was set manually, cpu clock speed remained the same, all volts remained constant.
- Cpu/NB frequency remained as close as possible no more than 50Mhz difference.
- Programs used were SiSandra, Everest, and Passmark.











There you have it, proof that 2000mhz ram is fastest when ran at it's intended speed.
Not only did it win all the tests, but it crushed the other speeds.

[GSKILL TECH]

Wait wait wait, where is the DDR3-2000 benchmark? haha

In any case, of course DDR3-2000 will be better/faster than DDR3-1600 or others.

DDR3-2000 = max bandwidth 16,000MB/s
DDR3-1600 = max bandwidth 12,800MB/s

This is fact. So obviously DDR3-2000 theoretically is much faster than DDR3-1600. But what people don't understand is that the other hardware must support it (bottlenecking). The perfect test is an overclocked P55 platform, that can support all memory frequencies. With that type of platform, you can easily tell the difference in memory bandwidth.

As far as latencies, depending on the memory, it may have certain "sweet spots", so even 7-7-7-24 and 7-8-7-24 may have a couple ns difference. Only testing and benchmarking can you find this, which is what max overclockers do.

magnus2882,

Great work, very informative, should be very helpful to other users.

Thank you
GSKILL TECH

[magnus2882]

thanks for the props!

i'll post the benchmark's again.

Note: i've even went down as far as 1066, but the results weren't worth pasting ...
2000Mhz best bandwith + latency

url=http://www.freeimagehosting.net/][/url]

[magnus2882]

why would I do that? Is 3 independant,reliable programs not enough to prove a point?

[GSKILL TECH]

I think it was just spam. The company probably has a person going around forums posting links to convince people to use their new programs. Well, he's banned now. =D

Thank you
GSKILL TECH

[mrcmtl]

Too bad that guy is banned, I would have asked him what's his NB frequency on those runs.

[GSKILL TECH]

magnus is not banned, it was some other guy that asked magnus to use other programs to test. He was promoting the other programs..

Thank you
GSKILL TECH

[CalBear96]

So would I have better performance running 1066MHz memory on an i7 640M whose FSB supports a max of 1066, or running 1333MHz that downclocks to 1066. Both sticks were the same price at Newegg, so I figured why not get the faster memory? If I will have better performance with the native 1066 I will exchange the memory. Thanks!

[genetix]

I don't think under 18000MB/s Read/Write it's even worth benchmarking at everest.

You'll get clean 18K at around 8-8-8-24 at around 1840-1870Mhz with any decensy at Uncore + CPU. depending of board of course how high it goes.

This is the slow version of usual clock at around 18500MB/s:


-edit-

hmm, mind remind everyone that above cannot be compared to 'AMD Phenom' benchmarks above as it's done with Intel's crappiest core I7-950 ever. ;P

[GSKILL TECH]

The faster one of course, considering the BIOS can downclock it. On a desktop it is no question since you can overclock the CPU. But for laptop it's all preset, so nothing you can really do about it.

Thank you
GSKILL TECH

So would I have better performance running 1066MHz memory on an i7 640M whose FSB supports a max of 1066, or running 1333MHz that downclocks to 1066. Both sticks were the same price at Newegg, so I figured why not get the faster memory? If I will have better performance with the native 1066 I will exchange the memory. Thanks!

[Beaser]

I'm glad to find this resource. The rules have certainly been rewritten when dealing with the I7series of processors.

I had always thought that the faster rated ram just passed more stringent tests within the series and that faster was always better. I'm building up a Ga-X58A-UD3R with an i7 950 & 12 gig of 1600mhz Ripjaw (12800CL9T-12GBRL) When I had a look at the POST and the bios and noticed the ram showing as 1066 and the cpu freq at 2.6 intead of 3.07.

I had a look around and tried the intelligent tweaker and the ram read as 1600mhz but the timing was too low. I restarted the sytem and rechecked and the timing seemed to correct them self(9-9-9-24).

I'm posting this so others are aware that this happens. After a while you get used to the quirks but any neophytes might have an anxiety attack.

The processor spec clearly calls for DDR3-800/1066 but you can't find any where I live and the 3 sticks of 1600 mhz was actually cheaper than the 1333.

I built an 1156 board up last month with an i5 and I've never had an issue with memory. Again I bought 1600mhz ram cuz it was cheaper.

Looks like a good product and at a reasonable price

[mgrobins]

which current generation CPU's benefit from higher clocked RAM (increasing bandwidth) as the CPU is overclocked?

Some in the past have not been bandwidth limited by the RAM and hence there was no point scaling your RAM up with the CPU speed.

Nice to see simple well presented results too .

[GSKILL TECH]

It's more the other way around, where the CPU overclocked better utilizes higher clocked RAM. Increased CPU and NB frequencies can drastically improve memory bandwidth.

Thank you
GSKILL TECH

[pstlouis]

Finally I can put this one to bed.

Some enthusiasts claim that because they can't run faster ram at the rated speed, that the same ram will actually perform better when downclocked - and at tighter timings. While i'm sure maybe SOME rams might behave this way, most simply will perform best when clocked at a higher frequency.

I've seen some benchmark's pulled off of Tom's hardware claiming that ram downclocked to 1600Mhz out performed the same modules clocked at 1800 and 2000. Let me just say I don't take much stock in benchmarks at tom's hardware- nor would I rely on them as gospel. I decided to run my own benchmarks, my own controlled environment, and use different software to confirm my results.

The Test Conditions:

- All benchmark's were performed right after a warm boot up.
- Each test was performed 3 times, and the best result was recorded.
- Bios was set manually, cpu clock speed remained the same, all volts remained constant.
- Cpu/NB frequency remained as close as possible no more than 50Mhz difference.
- Programs used were SiSandra, Everest, and Passmark.











There you have it, proof that 2000mhz ram is fastest when ran at it's intended speed.
Not only did it win all the tests, but it crushed the other speeds.

read the article then you will know what best speed are for the ram :

http://www.anandtech.com/show/4503/s...-the-best-ddr3

[GSKILL TECH]

Some times people rather not read, let alone do research, they just want somebody to tell them.

Thank you
GSKILL TECH