• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Benefit of using duel channel?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
wannaoc

wannaoc

Member
​
Joined
May 6, 2003
Location
Buried in UPS packages
What are they, my motherboards have had it I just never used it as I usually just have one stick. Will it help/hurt an o/c with a good sticks, or what? TIA.
 
Dual channel effectively removes any A-sync penalties you would otherwise incur by running your memory at a lower speed.

Before dual channel, if memory capped at a speed lower than the processor and motherboard, overclockers would run the lower FSB sync'd with the highest memory speed possible.

Because of how dual channel works, if your memory capped at something like 170mhz but your mobo and processor could do 218 mhz, you could set the memory ration to 80% and enjoy all the bandwidth you would have at a 170mhz sync'd rating, but have more overclocking options due to the capability (headroom) of your mobo / proc.

Basically, it really ONLY helps if your memory can't even come close to doing what your processor and motherboard can do.

And in the end, because of how FSB / multipliers interact, it only lets you get maybe 100 more mhz out of your processor, because it gives you more options in overclocking.

In the above example, here are two ways you could run your system:

10 x 218 @ 4/5 = 2180mhz cpu, 170mhz memory / 218mhz fsb
12.5 x 170 @ 1/1 = 2125mhz cpu, 170mhz memory/fsb

The above assumes that your proc and board can do 218 fsb stable, and the processor absolutely is not stable beyond 2180mhz.

Also, in the above example, system performance is basically identicle, with very small differences here and there.

My understanding is that dual channel does little to nothing at sync'd speeds.
 
InThrees is right on with his summary, at least I think so.

InThrees said:
My understanding is that dual channel does little to nothing at sync'd speeds.
Click to expand...

Regarding the gain at sync speeds of dual versus non-dual configurations, it is about 5+% as best I can determine. Results will vary I'm sure based on what you see from real applications.

Does it change what you buy?
- with a non-DC mobo you rather have 1 stick than 2,
- with DC take the 2 sticks over a single stick......why turn down 5+% performance gain you can get it by just by splitting the 1 stick into 2?

Be :cool:
 
dual pumped AMD's get the same benefit as dual and quad pumped intels. you get double the FSB. so if you're OC'd to 170Mhz and in single channel you bench at 2600MBps in dual channel you'll bench at like 5000MBps.

if dual DDR doesn't give us a performance increase than quad DDR doesn't help the intel guys.

the issue with the high FSB's and not seeing a great performance boost is because you're not pumping the memory to the same speeds.

you have:

CPU======northbridge=======memory
------^FSB------------^ mem spd*

dual or quad pumping the FSB doesn't dual or quad pump the other side of the northbridge so you're not going to get the amazing benefits you would think that come from high numbers.

this is my understanding of the situation. this is how corsair describes the relationship between CPU NB and memory and where the FSB and memory speeds come into play.

basically with these types of chipsets FSB isn't the bottleneck it's the other side of the northbridge.

this is also why intel would only support pc3200. because they wanted the memory speed to be a multiple of what the FSB would be.

*spd = in this asituation spd is speed not serial presence detect. i know that SPD is typically used to refer to the firmware on the memory. in this cause i used it to abbreviate speed so i could fit it. the dashes in between FSB and mem spd are there only because the post wouldn't save the spaces i put in. it's for formatting only not for representing a connection around the NB.
 
Last edited:
emericanchaos said:
basically with these types of chipsets FSB isn't the bottleneck it's the other side of the memory.
Click to expand...

If I understood your point emericanchaos (love that handle!), I would think I could observe that effect by setting me FSB to 100Mhz and crank the cpu multipler up by a factor of 2 and set the memory to 3/6 (Abit nomenclature), basically memory will be a factor of 2 more than the FSB.

I've done that and it doesn't appear to work. I measure the same memory MB/s as if I were running a memory speed of 100Mhz.

If I misunderstood, please straigthen me out. Goodness knows I need straighten from time to time.

Be :cool:
 
well i'm not 100% sure i'm correct. i'm throwing this out in hopes of being corrected where i'm wrong. if FSB accounts for the pathway on both sides of the northbridge than running dual and quad pumped FSB's would have a much larger performance boost than it does because the entire path would be pumped to double the memory speed. if that's truly the case AND we're not getting a great performance boost than you have to ask "where's the bottleneck?"
 
ok i'm back. i couldn't POST at 3/6 or 3/5. it gave a me long pulsed beeps than there was little beep beep boop boops while it had the long series of beeps. quite scarey. but i did manage to get my PC2100 to do 3/4 (woo hoo right) and my results differ a bit.

at 12*100 in 3/3 i did 1531 buffered and 1436 unbuffered.

at 12*100 in 3/4 i did 1556 buffered and 1472 unbuffered.

i ran twice in each setting and got the same both times.

so considering the slight difference made to the CPU-DRAM ratio and the slight improvement i'd say that changing the ratio like that can improve performance. and for people like me that don't have good memory we can still get high FSB by switching this to something like 4/3 or 5/3 or 5/4 etc because we're pumping the FSB higher while keeping the ram down. the question is what's better for performance. i'm going to reboot and go back to 12*100 3/3 and run 3dmark 2k1se. everyone says that it's FSB intensive so i'm going to see if i can get better results from it in something like 4/3 or 5/3 if it'll let me post.
 
I have been told that the advantage is minimal on amd cpu's, but on intel's that is a different story

EDIT: yay 3 stars :)
 
Andy71600 said:
I have been told that the advantage is minimal on amd cpu's, but on intel's that is a different story

EDIT: yay 3 stars :)
Click to expand...

your very correct...

Wow guys...dont be fooled by the gimics.

Lets do some very simple math here.

Ok intel "c" CPU.

800mhz, where does it get this?

Ok it has a FSB of 200x4=800mhz right?

so it's CPU is running a FSB of 800mhz and the actual FSB is 200, which means that 200mhz is going to the ram.

DDR ram makes that equal to 400mhz, dual channel makes that equal to 800mhz.

In an intel system since p4's are quad pumped it makes the ram and the CPU run in sync.

so summerize Intel platform here:

CPU = quad FSB
DDR=double FSB
Dual channel = double FSB

Dual channel DDR=quad FSB.


Ok but on an AMD this is completely differnt.

Lets say you overclock your AMD board to 200mhz FSB.

An XP athlon is "dual pumped" if you will, DDR FSB is what they call it i think.

So 200mhz FSB on an amd makes the CPU run at 400mhz.

DDR memory...remember it doubles it?

So 200mhz FSB going into DDR memory makes your memory run at 400mhz.

Look at that! it's already in sync.

So in theory running a 200mhz FSB on an amd dual channel system will run your memory at "800mhz" but it's really not doing much of anything, your AMD CPU simply cant "keep up" with your ram.

Now dont take this the wrong way, your performance doesnt drop, infact it goes up (VEYR VERY VERY VERY SLIGHTLY).

TO prove this please do some mem bandwidth tests...i have. Although this is coming from a guy who is running his AMD in dual channel.

Just wanted to clear that up.
 
InThrees said:
Dual channel effectively removes any A-sync penalties you would otherwise incur by running your memory at a lower speed.

Before dual channel, if memory capped at a speed lower than the processor and motherboard, overclockers would run the lower FSB sync'd with the highest memory speed possible.

Because of how dual channel works, if your memory capped at something like 170mhz but your mobo and processor could do 218 mhz, you could set the memory ration to 80% and enjoy all the bandwidth you would have at a 170mhz sync'd rating, but have more overclocking options due to the capability (headroom) of your mobo / proc.

Basically, it really ONLY helps if your memory can't even come close to doing what your processor and motherboard can do.

And in the end, because of how FSB / multipliers interact, it only lets you get maybe 100 more mhz out of your processor, because it gives you more options in overclocking.

In the above example, here are two ways you could run your system:

10 x 218 @ 4/5 = 2180mhz cpu, 170mhz memory / 218mhz fsb
12.5 x 170 @ 1/1 = 2125mhz cpu, 170mhz memory/fsb

The above assumes that your proc and board can do 218 fsb stable, and the processor absolutely is not stable beyond 2180mhz.

Also, in the above example, system performance is basically identicle, with very small differences here and there.

My understanding is that dual channel does little to nothing at sync'd speeds.
Click to expand...

Good read, i've never really looked at it that way...maybe i should try some tests with this :D
 
timmyqwest said:


Good read, i've never really looked at it that way...maybe i should try some tests with this :D
Click to expand...

My friend tried running 200mhz fsb 6/5 in dual channel for a 183mhz memory bus speed, or theoretically a 732mhz effective data rate. You would think that the bandwidth would not be limited by the memory bus, but it in fact was very much so- only 2400 MB/sec; not the 3200 that one would expect. Dual channel seemed to have done nothing.
 
Gautam said:


My friend tried running 200mhz fsb 6/5 in dual channel for a 183mhz memory bus speed, or theoretically a 732mhz effective data rate. You would think that the bandwidth would not be limited by the memory bus, but it in fact was very much so- only 2400 MB/sec; not the 3200 that one would expect. Dual channel seemed to have done nothing.
Click to expand...

Actually after i posted this and went into my bios i realized that there wouldnt really be any difference...but for someone with ram that limits them to a lower FSB i supose...it may help?

I dunno
 
ok yeah well i'm gonna have to hit the breaks on this whole thing. i received a "BIOS chip checksum error". it told me to run awdflash immeidately. i just unplugged the power cord and cycled it than powred again and here i am. don't think my BIOS is dead but i don't know. i thought the BIOS scrambling issues were dead on rev 2.0 boards.

anyways i can't get my old clock back. i used to run 12x160=1924. now i can't do more than 150. i tried leaving the memory alone and OC'ing only the multi and that's cool up to 12.5. haven't tried 13.

in any case i'm gonna spend some of my check i just got on some corsair and see if that fixes it. if not i'll be popping in the backup BIOS i got from excaliber for $2.

oh yeah thanks for the replies. i'm just trying to learn what this chipset does. i'm new with the whole cpu-dram ratio.
 
Last edited:
Dual channel improvement (if there is any) or difference over single channel for AMD nforce2 motherboard is around 2-3% overall taking into account of dual channel advantage minus the potential module mismatch which affects max FSB frequency, end result is only 2-3% (maybe 5%). Effective bandwidth improvement in P4 dual channel system can be as much as 50-60% over single channel, taking into account of memory channel actual efficiency.

Links:
Single Channel vs Dual Channel DDR Memory Module
Dual Channel, Nforce2, P4 and AMD FSB

hitechjb1 said:
Single Channel vs Dual Channel DDR Memory Module
...
Summary:

Dual channel or single channel mode in nforce2 mb is not that crucial for overall performance. The difference is only few % (say 2-3%) at most. Also single channel may let FSB to go a bit higher due to a smaller chance of potential dual dimm mismatch and memory controller stress at high FSB, I think. On the other hand, dual channel memory controller provides some performance advantage due to its intrinsic speculative caching capability. At this point, the little higher FSB from single channel offset the performance advantage of dual channel, and the two is about a tie, I think, for AMD mb. For some nforce2 mb that have integrated video which can benefit from twice the nforce2 memory bandwidth, since the bus between the video and the memory controller has 2x64 bit bus.

The max bandwdith between memory controller and CPU would be 2 x 8 x FSB = 16 FSB MB/s. x2 is because of DDR (data are transferred at both rising and falling edge of the FSB clock, x8 because of 8-byte bus or 64-bit bus). The effective bandwidth, taking into memory controller (~95% efficiency), would be around 15.2 FSB. E.g. FSB = 200 MHz, effective bandwidth ~ 3040 MB/s.

Dual channel makes a big difference for P4 dual channel mb though, due to quad pump data of P4 (or QDR). The max bandwidth for P4 dual channel is 4 x 8 x FSB = 32 FSB MB/s. The effective bandwidth, taking into memory controller overhead (~ 75% efficiency), would be around 24 FSB MB/s.

For single channel, max bandwidth = 16 FSB, effective bandwidth ~ 15.2 FSB. Hence the improvement of effective bandwidth of dual channel = (24 - 15.2)/15.2 = 58% for P4 dual channel system over single channel.

E.g. FSB = 200 MHz, effective bandwidth ~ 4800 MB/s, which is around 60% more than that of a nforce2 mb running same FSB 200 MHz.
E.g. running fsb:memory=5:4, with FSB=250, memory=200, effective bandwidth ~ 24 x 225 = 5400 MB/s.
Click to expand...

hitechjb1 said:
Dual Channel, Nforce2, P4 and AMD FSB
...
3. IMHO, running slower memory in ASYNC at 50-66-75% is much more price/performace effective than 100% SYNC. 50% memory cost will get to 10-15% of the max bandwidth. This approach has been used in P4 dual channel since its FSB is QDR (quad pump data) and there is no fast memory to match at that speed, and dual channel is the only way to fill up the system bandwidth.

For AMD DDR system, ppl who want absolute performance will just use the fastest memory to match the FSB and run them in SYNC to get the last 10% of memory bandwidth.

In summary, for AMD nforece2, the effective bandwidth running in SYNC mode = 0.95 x 2 x 8 x FSB = 15.2 FSB MB/s. E.g. FSB = 200 MHz, effective bandwidth = 3040 MB/s. FSB=225 MHz, effective bandwidth = 3420 MB/s.

For P4 dual channel, the effective bandwidth running fsb:memory=1:1 (SYNC mode)= 0.75 x 4 x 8 FSB = 24 FSB MB/s. E.g. FSB = 200 MHz, effective bandwidth = 4800 MB/s.
E.g running fsb:memory=5:4, with FSB=250, memory=200, effective bandwidth ~ 24 x 225 = 5400 MB/s.
Click to expand...
 
Last edited:
You must log in or register to reply here.

Similar threads

S
4x4GB DDR3 Overclock on AM3+
Replies
2
Views
337
Sologuy
S
SF101
AM5 6000MHZ Memory Timings Thread "SK Hynix"
Replies
16
Views
7K
SF101
SF101
A
newbie , ram for ga ep45-ud3p
Replies
4
Views
452
Sologuy
S
Dorito Bandit
Using Keyboard & Mouse With 2 Desktops
Replies
3
Views
312
Dorito Bandit
Dorito Bandit
ssjwizard
Mismatched RAM advise
Replies
5
Views
699
ssjwizard
ssjwizard
Back