Fast Loose RAM has the same bandwidth as Slow Tight RAM?

[Exahertz]

I don’t quite understand why fast loose RAM has the same “data rate” as slow tight RAM. What I mean by data rate is, when I multiply the frequency by the CAS latency I can come up with the same number from two different frequencies and latencies.
For example, if I have some PC-5300 with CAS4 @ 333MHz (4 x 3.0ns = 12 ns)
Verses if I have some higher PC-8000 with CAS6 @ 500MHz (6 x 2.0ns = 12 ns)
I know faster Ram with a tighter CAS is good, but why would I want to buy such high frequency RAM? What would be the benefits to the overall system?

I’m not sure if the term “data rate” is correct as the end result of the CAS multiplied by the frequency, please correct me if it isn’t! Thanks

[zanzabar]

thats not the timings that are getting supper high transfers u need like cas3 400mhz or cas 4 500


and it depends on what u are doing for benches low timings let the ram access more files faster but has slightly slower max bandwidth but since its got way less time to sit between files/banks it will bench better


but if ur streaming video into ram u might want the faster ram

[Exahertz]

wow, thats it? then whats the point of ddr3 :shrug:

[jimmyz]

wow, thats it? then whats the point of ddr3 :shrug:

lower power consumption and better fetching.

[CyberDruid]

I don’t quite understand why fast loose RAM has the same “date rate” as slow tight RAM. What I mean by data rate is, when I multiply the frequency by the CAS latency I can come up with the same number from two different frequencies and latencies.
For example, if I have some PC-5300 with CAS4 @ 333MHz (4 x 3.0ns = 12 ns)
Verses if I have some higher PC-8000 with CAS6 @ 500MHz (6 x 2.0ns = 12 ns)
I know faster Ram with a tighter CAS is good, but why would I want to buy such high frequency RAM? What would be the benefits to the overall system?

I’m not sure if the term “data rate” is correct as the end result of the CAS multiplied by the frequency, please correct me if it isn’t! Thanks

I'm no expert but from personal experience in general you are correct. However to be Xtreme is to be Fast and Tight:D

Why?

Each of the timings is about how long of a delay in terms of cycles so the less delay the faster the RAM can function and the higher the bandwidth.

If you increase the Frequency (number of cycle per unit of time) and the timings remain unchanged the bandwidth increases: i.e. more operations per unit of time so more data crunching.

If in order to function stable the delays for each operation are increased (higher timings and subtimings) then even though the RAM is clocking along it has to delay the operations to stay synched up..and works less efficiently.

Conversely if the RAM clocks along at a mediocre rate (by comparison) but has very little delay it gets more operations accomplished per unit of time irregadless of clock speed.

The ideal would be (obviously) to have very high clock speeds and very small delays on all the operations.

[nn_step]

in terms of sheer latency
DDR2-400 @ 3-3-3 is approximately equal to DDR2-800 @ 6-6-6
the DDR2-800 has double the available bandwidth.

[Exahertz]

So how is this ram (http://www.newegg.com/Product/Product.asp?Item=N82E16820145169) for bandwidth?
http://images10.newegg.com/NeweggImage/productimage/20-145-169-02.jpg
CORSAIR Dominator 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 1142 (PC2 9136) Dual Channel Kit

*Edit* Thanks Machinus

DDR2 1142 (PC2 9136)
5-5-5-15

1142MHz/2=571MHz (real clock speed)
5(1/571)=8.75ns

[zanzabar]

2 Gb Kit (2x1Gb) Cell Shock 4-4-4-12 1000Mhz
http://www.tankguys.com/product_info.php?cPath=46_144_176&products_id=1715

2 Gb Kit DDR2 (2x 1Gb) 5-5-5-15 1200Mhz
http://www.tankguys.com/product_info.php?cPath=46_144_185&products_id=1716



one of those 2 kits would be ur best options unless u want to order a set of cell shock red label custom

[Machinus]

You divide by the real clock speed, not the double clock speed. That would put the corsairs at 8.6ns.

[twilyth]

I had never thought to compare latency to clock speed so I thought this was a pretty sharp observation. I'm not sure I understood all of the responses, but this entry in Wikipedia (http://en.wikipedia.org/wiki/CAS_latency) helped me so I'm putting it out there.

Comparing between clock speeds gets trickier. CAS Latency only specifies the delay between the request and the first bit. The clock speed specifies the latency between bits. Thus, when reading bursts of data, a higher clock speed can be faster in practice, even with a worse CAS Latency. For example, consider a 133 MHz CL3 device (7.5 ns per cycle, 3 cycles request latency) versus a 100 MHz CL2 device (10.0 ns per cycle, 2 cycles request latency). The first bit would be available after 22.5 ns (7.5 ns * 3) on the CL3 device and after 20.0 ns (10.0 ns * 2) on the CL2 device, demonstrating the benefit of a lower CAS latency. However when reading a burst of even 4 bits, the higher clock speed wins: 45.0 ns (7.5 * 3 latency + 7.5 * 3 bits after the first) versus 50.0 ns (10.0 * 2 latency + 10.0 * 3 bits after the first).

The example is pretty dated, but I think it gets the idea across.

[zanzabar]

before the people with the cheater x9xxx intel not public chips on 3d mark the top was a system with cas3 925mhz or something around there

[Exahertz]

2 Gb Kit (2x1Gb) Cell Shock 4-4-4-12 1000Mhz
http://www.tankguys.com/product_info.php?cPath=46_144_176&products_id=1715

2 Gb Kit DDR2 (2x 1Gb) 5-5-5-15 1200Mhz
http://www.tankguys.com/product_info.php?cPath=46_144_185&products_id=1716



one of those 2 kits would be ur best options unless u want to order a set of cell shock red label custom

That is some pretty wild ram! Is 1200MHz the real clock speed like Machinus said? if so, that puts that stuff down in the 4.16ns range.

[zanzabar]

those are a much better deal than the corsair arnt they, but when u get into $300 for a 2gb kit i would lean to getting crucial ddr3 since they use micron across the board for ddr2/3