memory 1t or 2t?

Zantal · Apr 12, 2009

ok so my memory atm is working at 2t. some ppl say it's better to make them work at 1t.

but first i don't know the pros and cons of this settings.

if someone could please explain me i would really appreciate it.

(btw i have an asus p6t deluxe and i didn't find such option in the bios)

BG-0 · Apr 12, 2009

Depends on your RAM. Theoretically 1T is better.
The 2T/1T/0T option should be available in the memory settings, below the timings. (CAS, etc) Shuffle through the BIOS and you should find it. After you have changed, do a few memory benchmarks, change back, benchmark again. Whichever works better for you, keep that.

Zantal · Apr 12, 2009

thanks. gonna check now

redduc900 · Apr 12, 2009

In the P6T BIOS the Command Rate option is called DRAM Timing Mode (under AI Tweaker--> DRAM Timing Control). The settings 1N, 2N and 3N coresspond to 1T, 2T, and 3T CRs respectively.

Badbonji · Apr 12, 2009

If you have 3 sticks or less, than 1T should be fine. 2T is generally for 4-6 sticks, with ddr2 and ddr3 as there is more stress on the memory controller with more sticks.

Zantal · Apr 12, 2009

worked good.

yepp i have 3 sticks only (3 * 2gb) so i set it to 1N, 1gb bandwidth more . . . although i think the cpu won't be able to use it all anyway =)

redduc900 · Apr 12, 2009

Zantal said:
1gb bandwidth more . . .although i think the cpu won't be able to use it all anyway =)

Explain please.

Badbonji · Apr 13, 2009

Explain what? The lower command rate will increase bandwidth, as to the cpu needing it is another question. I get over 30Gb/sec of bandwidth in Sandra, and I am thinking the CPU doesn't really need it all in most real life situations

Zantal · Apr 13, 2009

i have 26gb/s @ 1.45 ghz memory, but i remember an article showing that real apps don't need such huge amounts. difference between dual channel and triple channel bandwidth didn't increase overall performance.
but i like to have the best things for my pc anyway =)

redduc900 · Apr 13, 2009

Badbonji said:
Explain what? The lower command rate will increase bandwidth, as to the cpu needing it is another question. I get over 30Gb/sec of bandwidth in Sandra, and I am thinking the CPU doesn't really need it all in most real life situations

Well my question was directed to the OP, but since you asked... how is the CPU affected by the command rate? You stated:

Badbonji said:
I am thinking the CPU doesn't really need it all in most real life situations

... and the OP stated:

Zantal said:
i think the cpu won't be able to use it all anyway

... so I'm curious to know how the memory bandwidth (in effect a lower or higher command rate) affects the CPU.

Zantal · Apr 13, 2009

redduc900 said:
... so I'm curious to know how the memory bandwidth (in effect a lower or higher command rate) affects the CPU.

well try thinking it this way

consider the cpu a Car factory and the mem bandwidth the steel supplier

imagine now the steel supplier is very very fast and brings tons of steel to the car factory.

but the factory can't produce so many cars per day, most of the supplier capacity is wasted.

i hope i am not confused about it, but i tried to explain it this way.

and

The Command Rate is the time between the chip select signal and when commands can be given to the RAM

so if set to 1 it will go 1 clock before command and so on till 3

Zerix01 · Apr 13, 2009

Zantal said:
well try thinking it this way

consider the cpu a Car factory and the mem bandwidth the steel supplier

imagine now the steel supplier is very very fast and brings tons of steel to the car factory.

but the factory can't produce so many cars per day, most of the supplier capacity is wasted.

i hope i am not confused about it, but i tried to explain it this way.

My understanding was that if data is not in a register or in the CPU cache then the CPU will be waiting until the data is fetched from either the RAM or storage device. Any waiting means waisted CPU cycles. Any added bandwidth should be usable by the CPU in most tasks. But whether or not the change is noticeable in the given application is another story.

Archer0915 · Apr 13, 2009

So if you dont need the bandwidth then lower the speed ant titen timings it will give better response.

Hipcrostino · Apr 13, 2009

DDR and DDR3 will run at 1T (or 1N) command rate just fine, or at least, thats what you should aim for. It speeds things up quite a bit in benchmarks.

DDR2 will almost always refuse to run at 1T, and should be run at 2T. only the hardcore guys use 1T on DDR2.

Thats about it. But if you are unsure, or all else fails, 2T is always a safe bet.

Zantal · Apr 14, 2009

Zerix01 said:
My understanding was that if data is not in a register or in the CPU cache then the CPU will be waiting until the data is fetched from either the RAM or storage device. Any waiting means waisted CPU cycles. Any added bandwidth should be usable by the CPU in most tasks. But whether or not the change is noticeable in the given application is another story.

maybe it is. but i don't see the point of all this race to teh bandwidth going on.

we aren't working on servers

which needs tons of that

JackNSally · Apr 16, 2009

1t should give better latency making the system feel more responsive. Whether the cpu really needs the added bandwidth is a different story.

KillrBuckeye · Apr 16, 2009

JackNSally said:
1t should give better latency making the system feel more responsive. Whether the cpu really needs the added bandwidth is a different story.

Are you really sure? Whether it's 1T vs. 2T command or tightening up of the memory timings, I have never noticed a difference in the "responsiveness" of a system. I've had people tell me that applications open faster and respond more quickly when using 1T vs. 2T (or tighter timings), and based on my personal experience I just don't believe it. If there were some way of actually measuring this sort of thing, I would love to see some data, but I can't help but think this is purely a placebo effect.

Jmarsh667 · May 14, 2010

1N

KillrBuckeye said:
Are you really sure? Whether it's 1T vs. 2T command or tightening up of the memory timings, I have never noticed a difference in the "responsiveness" of a system. I've had people tell me that applications open faster and respond more quickly when using 1T vs. 2T (or tighter timings), and based on my personal experience I just don't believe it. If there were some way of actually measuring this sort of thing, I would love to see some data, but I can't help but think this is purely a placebo effect.

Your right no one will notice pico seconds your talking a trillionth of a second maybe a benchmark will see a very small difference but a person will not.

Marshmallow64 · May 14, 2010

For me I play around with it sometimes when I overclock, I am on 2T right now for stability.

wingman99 · May 25, 2010

KillrBuckeye said:
Are you really sure? Whether it's 1T vs. 2T command or tightening up of the memory timings, I have never noticed a difference in the "responsiveness" of a system. I've had people tell me that applications open faster and respond more quickly when using 1T vs. 2T (or tighter timings), and based on my personal experience I just don't believe it. If there were some way of actually measuring this sort of thing, I would love to see some data, but I can't help but think this is purely a placebo effect.

+1 it is the placebo effect. The speed of the modern DDR2 and DDR3 memory bandwidth is faster than the MCH and CPU can handle and respond with real world programs, especially when the CPU needs small chucks of data from memory.

Where as a video card needs DDR5 because your moving and spinning in a game it loads the map data fast using all the bandwidth efficiently hence there is less complications in retrieving the memory sinces the GPU just works with graphics.

memory 1t or 2t?

Member

Member

Member

Inactive Moderator

Member

Member

Inactive Moderator

Member

Member

Inactive Moderator

Member

Member

"The Expert"

Member

Member

Member

Senior Member

New Member

Member

Member

Similar threads