|
RAM speeds and how they relate to production?
|
View this Thread in Original format
| Stephen Wiley |
Just wondering what would be better suited for production (with ableton live) here. Using windows 7 64bit
DDR3-1600 (timings unknown)
DDR3-1333 (8-8-8-24)
This is from the manufacturer (G.Skill)
his memory package should work just fine. For DDR3-1600, simply set the memory frequency, timings, and voltage in BIOS. If you experience slight instability, raise CPU-NB Voltage +0.1V and that should stabilize it.
For DDR3-1333, use 8-8-8-24 timings and all should be just as fast as DDR3-1600.
Also if anybody here knows how RAM effects production and what is most important/least important, I'm sure many more on this board would be interested to hear. (Stability is obviously #1) |
|
|
| evo8 |
I dont think you will see much of a real world difference in timings on a music production machine but i could be wrong
You could have a look at some of the benchmarking sites like toms hardware but they always tend to focus on graphics and gameplay. I was under the impression that faster RAM and timings were for overclocking purposes
My RAM is "Corsair Value Select 2GB (1x2GB) DDR3 PC3-10666C9 1333MHz Memory Module" (6 gig)
...timings are 9-9-9-24 which is standard enough i think
Im on Live 8 on win 64 bit also, takes everything i throw at it really |
|
|
| DJ RANN |
IME, once you reach the top price brackets of RAM, the difference in rice gets you minimal increase in performance.
The 1333 can be easily OC'd to get the same performance as the 1600 without any issues but you have to make sure your processor can handle that spec, for instance from what i understand the some of i-5's can only make use of 1333 so OC can be essentially useless for those lower models.
The upper models and i-7's however can utilise 1600 (even though they only recommend 1333) so OC'ing them could be a cheap way of getting good performance.
So the immediate answer could be to go for the cheaper RAM but.....
You may also want to factor in CAS latency, it refers to how quickly the RAM can respond in certain ways - a low figure is better. For example, that RAM you linked to has a CAS latency of 9 cycles at 1333 MHz (6.75ns). However, the higher end memory, such as the Corsair Dominator 6GB kit has a CAS latency of 8 cycles at 1600 MHz (5ns). As you can see, the Dominator responds faster - in 5 nanoseconds instead of 6.75. This may mean all in terms of performance for most tasks but audio, especially very large amounts of sample data could be affected by this. |
|
|
| Aesthetic |
Dont forget if you're not overclocking, getting performance RAM is kind of pointless, look up factors in relation to bus speeds and RAM speeds, but any RAM running in dual channel has the capability to run at double the data rate
*edit* shows how "oldschool" my core 2 duo is now.. the core-iX family dont even use an FSB :) |
|
|
| Stephen Wiley |
I ordered in conjuction with it an AMD 3.4ghz Quadcore Black Edition (3.2 initially) and am using a Gigabyta 880-GA UD3H motherboard along with an after market heatsink. Probably won't be doing any overclocking unless I max out my CPU in Ableton, but from what I have been reading it's going to be hard to do that, but I got a nice after cooler to preserve processor life and also give me the option to overclock if I do end up needing to.
I used to avoid checking out synths like DCAM Synth Squad because it would eat 30% of my processor on its own. DSP Discovery when set at high voices would max the CPU out in just one instance. Juno V2 from Arturia was reaching 30%, etc. So there were synths out there I was having to avoid with my 2.4ghz core2duo that now I am not going to be afraid of. |
|
|
| Rodri Santos |
| Producing it's mostly processor demanding, if there is a lot of cpu usage appears a groovy noise. As for RAM the DAW can use up to 1Gb aprox. If you have 2 GB DDR3 no matter the speed... it's fine. |
|
|
| DJ RANN |
| quote: | Originally posted by Rodri Santos
Producing it's mostly processor demanding, if there is a lot of cpu usage appears a groovy noise. As for RAM the DAW can use up to 1Gb aprox. If you have 2 GB DDR3 no matter the speed... it's fine. |
Dude, who's ing alt are you?
Most serious producers I know easily go through 2gb of ram just opening their sampler(s). Speak to anyone that uses real strings and they'll laugh you out of the room - upwards of 6gb is the start point. |
|
|
| DigiNut |
Production work tends to be limited, in order of importance, by:
a) CPU speed;
b) Total RAM;
c) Disk speeds (for multi-track recording and DFD samplers in low-memory situations).
The memory bandwidth and/or latency in your machine determines the speed at which data can be loaded into and out of RAM. Samplers really only have to load a single set of data one time, since the samples themselves never change. Synthesizers, in theory, do perform a lot of memory operations, but in most cases there's so little data that it all fits into the CPU cache anyway, and the occasional cache miss isn't going to seriously hurt.
I'm sorry I don't have reams of benchmarks for this, but for years now I've held the opinion that "premium" models of RAM (i.e. with 10% faster speeds) are basically a sham, costing consumers a lot of money without delivering any noticeable practical value. That's not to say you should buy the cheapest off-brand memory you can find, since the reliability varies quite a bit and you really don't want flakey RAM - but as far as speeds go, don't waste your money.
Pick a manufacturer you trust, pick the memory size you need (as large as you can afford!) and buy the lowest-end model from that manufacturer in that size that's still in production and that your mobo will accept. |
|
|
| DJ RANN |
| Good info but would you think this applies for synths such as Omnisphere which are both sample and CPU intensive? |
|
|
| kitphillips |
| Theres definately diminishing returns but DFD samplers these days would probably benefit from faster RAM I would have thought... They load many samples from disk into ram constantly as you need them. Of course, I suppose the limiting factor there is how fast your disk can load data into the RAM, so maybe thats why it makes no difference. |
|
|
| tehlord |
Although these tests aren't specifically DAW related, they do indicated that it's the amount of RAM and not the speed of the RAM that's the biggest factor in performance. I think the difference in performance between the fastest and slowest RAM is surprisingly small.
http://www.extremetech.com/article2...,2328813,00.asp |
|
|
| DjStephenWiley |
well i got the new computer up and going
4gb of g.skill ripjaws
amd quadcore 3.4ghz
gigabyte motherboard 880GA UD3H
Katana3 heatsink/fan with arctic silver
So far, so good. Very pleased with the G.Skill. Popped the sticks in, BIOS detected DDR3 1600 with the proper timings, and not a problem since. Will def. be getting 2 more sticks and I recommend anyone out there to check these out. only $99 on newegg right now for 4gb. cool looking RAM too.
thanks for all the replies ! |
|
|
|
|
