who uses a 1:1 ram ratio?

hypocrisyforever

Distinguished
Mar 30, 2008
155
4
18,715
So, I've been overclocking for awhile, and I don't understand how anyone can use a 1:1 ram ratio. I have pc8000 ram rated to run at 1,000mhz, which is pretty fast and even at that, I can't run my stock cpu frequencies at 1:1 without them blowing my ram rating out of the water......Am I missing something?
 
I had the same issue with my latest build so I just did 400Mhz FSB then did 1:1 ratio for my DDR2 800MHz ram. Running 1:1 or not, the difference in performance won't be large. It may or may not be worth the time to get it stable for such little performance benefit.
 
See, that makes no sense to me, care to explain? If I put my fsb to 400 with my quad pumped ram, a 1:1 ram ratio puts me at 1600 on my ram?.....ie, 600mhz over my 1000 cap.....explain?
 
DDR2 667 = FSB 333.5. DDR2 800 = FSB 400 DDR2 1000 = FSB 500 These are 1:1 ratio, you are unlikely to be getting a successful overclock with that board with a 500 FSB and a 1:1 ratio. However, given that the chipset on that board is the 680i, I would hesitate to overclock at all, given the number of issues that users have encountered with it, notably with data corruption during overclocking.

http://www.overclock.net/intel-motherboards/167489-asus-p5n32-e-overclocking-guide.html try this for some extra info if you want.
 
You're likely thinking of the Intel processors, they are 'quad pumped'

Your PC2-8000 RAM is aka DDR2-1000, divide that in half ('double data rate') and you arrive at 500mhz for your memory.

This means you can overclock your processor's FSB up until the 500mhz point (which is the 1:1) before having to overclock your RAM to get a higher processor overclock.

Your profile says you're using a Q6700 (2.66ghz stock), which has a multiplier of x10, and runs at 1066fsb --> this 1066 fsb is what is 'quad pumped', so divide by 4 and you get 266fsb. x10 = 2.66ghz stock speed.

Now, what your memory allows you to do is bring that 266 up to 500, and x10 = 5.0ghz, which is theoretically what your processor could overclock to....but you'll be seeing smoke long before you get anywhere near that of course :)

I hope that clears it up.. I'm pretty tired at the moment so it may require reading once or thrice haha

Cheers!
 
yeah, so RAM is "Dual pumped" cos its Double Data Rate.... but then if you have two sticks in "dual Channel" configuration, which is supposed to double the bandwidth....wouldn't that make it "quad pumped"??
 


To follow up on this, if you use a 7X multiplier with 500FSB, you will be looking at a CPU clock of 3.5GHz, which is possible with that CPU with good cooling. 8X400 = 3.2GHz, certainly within your ram and processor's capability, and with the benefit of probably being able to run tighter ram timings than the higher FSB would allow.
 


Hmm well the simple/short answer is basically no, not when it comes to talking about the bus speed of the memory itself. You definitely want to run your sticks in dual channel mode because it is much more efficient; that's why if say you have 3gb of memory, it's better to have 2x1gb and 2x512mb so they can operate in dual channel. But that doesn't actually affect the bus speeds... unless one of your sticks is a different brand/model and has a slower speed, then the other memory will down-clock themselves to match.

Thats about as good an explanation as I can give tonight, I'm off to bed!

Oh and Saunders makes a good point: You can 'tighten up' or lower your memory's timings (from 5-5-5-15 to 4-4-4-12, for example) if you aren't making full use of the fsb.
 
Your profile says you're using a Q6700 (2.66ghz stock), which has a multiplier of x10, and runs at 1066fsb --> this 1066 fsb is what is 'quad pumped', so divide by 4 and you get 266fsb. x10 = 2.66ghz stock speed.

So, what does this 1066 QDR mean? like, what does it do for the system, what job does it perform? that is the only thing lefta bout the whole FSB thing that i don't understand... :kaola:
 
I guess to make it simpler you could look at it a little differently. The reason people use a 1:1 memory ratio is because there is little to no real world benefits other wise.

If your cpu fsb is 1066mhz and yoru ram total speed is 2000mhz, the system can only provide data to the cpu at 1066mhz. basically the extra speed is wasted.

now if you O.C. your fsb up to your ram speed you will notice drastic gains in performance not just from the increase in clock speed from the cpu but from the increase in memory performance.

Hope that helps to answer your orginal question "why do people use 1:1 ratio"
 
the front side bus is the data transfer bus that carries data from your CPU to the NorthBridge...

The bandwidth or maximum theoretical throughput of the front side bus is determined by the product of the width of its data path, its clock frequency (cycles per second) and the number of data transfers it performs per clock cycle.

so just think of it, as a freeway lane between your CPU and NB, the bigger it is, the faster and more data can be sent through. and the NB is connected to the AGP or PCI-e slots, along with the Ram slots, so that is how the information is getting to your CPU.

the frequency of the CPU is determined by the multiplier and the FSB... example would be the Q6700 = 266x10 = 2.66ghz or 2667 megahertz

an easy way to remember the ram and FSB ratio...for 1:1 is just to divide the FSB by 2 and that'll be the # in mhz that your ram needs to be to run in 1:1. of course people above me explained the details, but if its hard to understand, just divide by 2.

ex. FSB 1600/2 = 800mhz. so for a FSB of 1600, you would need a ram capable of 800mhz in order to run synchronous with it.

DDR2 ram is "doubled pumped" ...
CPU FSB is "quad pumped" ...

lets use the Q6700 again

266 mhz (FSB) x 10 (multiplier) = 2.66 ghz (CPU frequency)

266 x 4 (quad pumped) = 1066mhz = your FSB frequency

so for a 1066mhz FSB frequency speed, you would need ram that is 533mhz or above to run it in 1:1 ratio

if you OC 266mhz to 333mhz you would need 667mhz ram

333mhz x 4 = 1333...and as i mentioned before, 1333/2= 667, so dividing by 2 is an easier way to remember it.

hope that cleared things up :)
 
Couldn't have said it better myself.

And people go around thinking you can shove 533 (1066)Mhz worth of Data through a 400(800) Mhz FSB.

Won't quite work!

--Lupi
 
He managed to pack that in there, and I was replying to PsyKhiqZero!

hehehe!

Man, I tried to make the people over at extreme system forum under stand about VID, and wow, they all wanted to kill me!

It's like no one wanted to learn!

You should really try it.

They still wonder why one q6600 gets 3.6 at 1.300 volts, and others get no where near that the the voltages!

I told em it was prolly because their VID determines their starting voltages, and they about had a break down.

It's common knowledge over here, but it has yet to be proven over there.

Crazy!

And I dont feel like proving it all again!

Once here was enough for me! To much typing!

--Lupi
 


my understanding is that dual channeling doubles the bus width and not the clock speed , more data per clock rather than more clocks

the result is the same as far as transfer rates are concerned
 
It's common knowledge over here, but it has yet to be proven over there.
can ya post a URL to that post so i can understand what VID is lol? i dont no a single thing about it...not even wat it stands for lol :na:

my understanding is that dual channeling doubles the bus width and not the clock speed , more data per clock rather than more clocks
well the clock "speed" just means how quickly the RAM executes a clock cycle doesn't it? doesn't mean more clocks as such... :non:

kind of technical subject aii... :ouch: 🙁 😗 :cry: :kaola: :non: :pfff: :bounce:

harharhar.... f34r my 133t 5k!11z...
 
VID, the Voltage ID of your processor.

The assigned VID, IE, BIOS VCore when on Auto and stock speeds.

Obviously from there you can use logic to see its uses.

The lower you start, the lower you get to end.

For instance, the q6600 series. The VIDs range from 1.2000 to 1.3250. It goes in .0125 increments. 1.2000, 1.2125, 1.2250, 1.2375, etc.

I have a VID of both the low and high end. The low end gets 3.6 @ 1.3 volts, while the 1.3250 gets the same speed at 1.440 volts loaded.

Big difference.

--Lupi
 





ok, thanks for the help....definitely makes sense to me now. I was misled by my board because if i set the ram ratio to 1:1, it shows my "actual memory frequency" as exactly whatever my fsb is....not half. This is what confused me, but I can tell you it makes alot more sense that the ram mhz is 1/2 the fsb when in 1:1. Still not sure why my board shows it as the full fsb speed....but hell, you answered my question ;-). Thanks again guys.
 
Ah, but wait... there's more! :pt1cable: Let's consider some of the additional variables involved in the memory big picture: 😱

A ratio of 1:1 provides the best level of stability, since the memory controller, which is an integral part of the northbridge chipset for Intel processors, does not need to translate data flow across the FSB between the memory modules and the processor(s). Also, since memory and processor FSB clocks are synchronous at 1:1, (400:400 or DDR 800), there is no additional latency introduced.

If a minimal ratio of 4:5 (400:500 or DDR 1000) is used, then the resulting increase in memory frequency is effectivey cancelled out by the latency introduced in translation across the FSB between memory and processor clocks, and no increase in memory performance can be noticably detected in benchmarks. Also, asynchronous or mismatched clocks create an element of potential instability within the memory controller, so depending on the chipset, an increase in northbridge and memory voltage is required for stability, which results in more heat, and less FSB overclock ceiling.

If a more aggressive ratio of 2:3 (400:600 or DDR 1200) is used, then the increase in memory frequency can marginally overcome the latency introduced in translation across the FSB between memory and processor clocks, resulting in a marginal increase in memory performance, which typically yields an increase in memory benchmarks of 2 to 3%, and is relatively negligible in terms of overall system performance.

In the case of DDR3, where a ratio of 1:2 (400:800 or DDR 1600) or 2:5 (400:1000 or DDR 2000) is used, even with the tightest timings, an increase in memory benchmarks of only 3 to 4% is yielded over DDR2 800, which once again, is relatively negligible in terms of overall system performance. It's also noteworthy to consider that DDR2 memory timings of 4-4-4-12 compared to 5-5-5-15, will yield an increase in memory benchmarks of less than 2%. However, for those of us who have the need for speed, we'll take whatever we can tweak.

Additionally, at equal specifications, 4 memory modules offer less FSB overclock ceiling than 2 modules, since more northbridge and memory voltage is required to maintain stability, and 4 slots require twice the current of 2 slots, again resulting in more heat, which typically is pulled into the CPU cooler, where it can increase processor temperatures by a few unwanted degrees.

I hope this helps to provides a greater degree of perspective (no pun intended). 😀

Comp :sol:

Incidentally, I run my rig at 1:1.
 
Computronix
Nicely explained, maybe this will help some people understand that memory overclocking yields minimal benefits. I have seen some objective testing where DDR2 667 yielded better net performance than 1066 or 1333, due in part to decreased latency and tighter timings.
 
yeah computronix, that yielded some more light on the situation...
i kind of knew most of it except the "4 memory modules" thing.... that's clinched it for me now!! i was thinking of getting two kits of (2gb - 2x1GB sticks) for either an evga 750i ftw or some P45 board..... now i'll just grab 1 kit of 4 GB (2x2GB sticks)...

thanks again
V
 
Is that Computronix? A non temp thread? :)

Hehe!

See, I recommend Real Temp and Core Temp now!

And core temp, the newest version is foiled by EIST if it is enabled.

Hmmm, that P45 chipset yields such high FSBs... Do you think that intel corrected the signal degradation of splitting the memory controller signals into two channels?

You think they would, so we can use all 4 to the same degree!

Integrated Memory controllers! (will that even help?)

Will you have to over clock your Integrated memory controller to get more than 1066 out of it? (Joking!)

I feel a bit deranged...

--Lupi