Just now, Ray Proudfoot said:

If that is the case then I have been given conflicting advice here on AvSim. 🙄

 

By who?

 

6 minutes ago, martin-w said:

By who?

Read Bob's third para in this reply. That would appear to conflict with your link to Crucial.

 

Yeah I'm not sure I would agree with that. The "real measure of performance" isn't just true latency, it's frequency too. And BOTH play a part. In the example you gave the true latency between your two examples is barely any different, but the frequency is way higher. I would agree with Crucial, after all they are the experts.

I wouldn't be too concerned Ray. Tests here on Avsim in FSX demonstrated only a few percent difference, no big deal, and that was (IIRC ) jumping from 3000 to 3800 MHz. 

Again, I'm not a P3D expert though, re how RAM impacts it. But I have no doubt that you will be happy with your 3000 MHz RAM, and as I say, it can be nudged up in frequency rather easily. 

 

"The real measure of RAM performance is true latency"

I would rephrase his comment as follows... The real measure of latency is "true latency" not CAS latency. 

He also said... "Yup, no faster at all, in fact if you take the heat spreader off, don't be surprised to find the same memory ICs on the DIMM."

 

Well yes they will be, they are speed binned. 

Edited November 1, 20187 yr by martin-w

9 minutes ago, Ray Proudfoot said:

Read Bob's third para in this reply. That would appear to conflict with your link to Crucial.

 

I don't think it does conflict.  I loaded the Crucial examples into Excel and the numbers match Bob's formula.  Speed and latency both have to be included in determining relative performance.  The Crucial page states that CL alone is misleading because it's only half the equation.  Then, in their last paragraph they state that speed is more important, which is still only half the equation.  So their conclusion doesn't match their own article.

i went with 3200cas14 for financial reasons, but had I had no limit on spending I would have gone higher.  Mine are GSkill, just running at XMP profile, havent tried running them higher yet...

3 hours ago, Ray Proudfoot said:

That's interesting. I've just ordered a new PC based on an i7-8086K. The company said they can overclock up to 5.0Ghz but I wonder if they might be able to get more. My memory is 'slower' than yours but when I asked them if 'faster' memory would help overall performance they said not so I went with G.Skill Trident Z RGB 32GB DDR4 3000MHz.

I'll report back when I get the machine in about 10 days.

Ordered same, Ray, from Scan a couple of months ago..

Good to see you at Cosford once more

5 minutes ago, LRW said:

Then, in their last paragraph they state that speed is more important, which is still only half the equation.  So their conclusion doesn't match their own article.

 

I think they were saying that "true latency" has hardly changed, only CAS latency has changed, thus speed being the important factor. 

21 minutes ago, martin-w said:

Yeah I'm not sure I would agree with that. The "real measure of performance" isn't just true latency, it's frequency too. And BOTH play a part. In the example you gave the latency between your two examples is barely any different, but the frequency is way higher. I would agree with Crucial, after all they are the experts.

I wouldn't be too concerned Ray. Tests here on Avsim in FSX demonstrated only a few percent difference, no big deal, and that was (IIRC ) jumping from 3000 to 3800 MHz. 

Again, I'm not a P3D expert though, re how RAM impacts it. But I have no doubt that you will be happy with your 3000 MHz RAM, and as I say, it can be nudged up in frequency rather easily. 

 

"The real measure of RAM performance is true latency"

I would rephrase his comment as follows... The real measure of latency is "true latency" not CAS latency. 

He also said... "Yup, no faster at all, in fact if you take the heat spreader off, don't be surprised to find the same memory ICs on the DIMM."

 

Well yes they will be, they are speed binned. 

Given it's a mathematical formula you would expect it to be either correct or incorrect. No half-way house.  There is very little difference in actual speed using that formula which presumably was thought out scientifically.

And as you say the difference is very small so I'm not going to lose any sleep on it. But it would be nice if everyone on AvSim could agree on the best way of selecting RAM.

If Bob reads this topic he may care to discuss that.

14 minutes ago, martin-w said:

 

I think they were saying that "true latency" has hardly changed, only CAS latency has changed, thus speed being the important factor. 

Yes, they did state that.  In fact, it was the central point of the article.  But the sentence stating the speed is more important than latency struck me like marketing had to stick their thumb in it at the last moment.  I first saw this article on their site a few years ago and it's still the best consumer explanation of memory performance that I've seen.

26 minutes ago, LRW said:

I don't think it does conflict.  I loaded the Crucial examples into Excel and the numbers match Bob's formula.  Speed and latency both have to be included in determining relative performance.  The Crucial page states that CL alone is misleading because it's only half the equation.  Then, in their last paragraph they state that speed is more important, which is still only half the equation.  So their conclusion doesn't match their own article.

It seems Crucial are just confusing people. Bob didn't just take CL or speed on their own. They formed the basis of the formula he provided.

Maybe that link to Crucial needs removing as it doesn't seem particularly helpful.

Quote

speed is more important than latency struck me like marketing had to stick their thumb in it at the last moment.

 

I don't blame you being suspicious, but I think in the context they were referring to it makes sense. 

30 minutes ago, vc10man said:

Ordered same, Ray, from Scan a couple of months ago..

Good to see you at Cosford once more

Are you enjoying your new toy Rick? Nice to see you too. 👍

36 minutes ago, Ray Proudfoot said:

Are you enjoying your new toy Rick? Nice to see you too. 👍

Waiting to pick Martin's knowledge base for a suitable RAM+MoBo. CPU air cooler is fixed.as per Martin's.

1 hour ago, martin-w said:

 

I don't blame you being suspicious, but I think in the context they were referring to it makes sense. 

Of course, having said all of that, I have the TridentZ 4400 CL18 to put in my Maximus XI Code whenever I receive my unicorn..sorry.. i9-9900K (I get those confused as they have the  same availability).  The total latency of the 4400 is lower than that of the TridentZ's with lower frequencies.  I hope it boots when I turn on XMP.

Which planes do you use?