2400 or 2600 MHz RAM for Ivy?

[Capt Lalle 0]

Hi,

 

does it make a noticeable difference between 2400 MHz or 2600 MHz RAM for FSX on an oced Ivy platform?

 

I'm thinking of buying GSkill Trident X 8 GB 2600 or GSkill Trident X 16 GB 2400 RAM. This will be a FSX only platform, therefore it's not really necessary to take 16 GB of RAM I think.

 

Thanks in advance

[dazz 69]

2% give or take

[Noel 4,486]

2% give or take

 

Though it would be helpful to understand what dazz meant by '2%' I would have to say the answer, despite having zero experience empirically with the question at hand and pardon me for chiming in, would be...no, it will make no noticeable difference! Now, when I do my big upgrade to Haswell next year, will I put the fastest RAM available into the build, why...yes, of course! At least you can say you did everything possible to help the old code run smoothly, even if it didn't really matter!

 

Happy New Year!

[TechguyMaxC 516]

It's not a huge gain, I went from 2133 to 2600 and saw maybe 1-2 FPS gain on average IIRC. Of course, 2400->2600 is an even smaller gain, I'd say 2400 is about where Ivy stops scaling nicely. That being said, if you're all about maximum performance and the price difference isn't huge, go for it!

[dazz 69]

I'd say 2400 is about where Ivy stops scaling

 

I'm pretty sure Hasse's gonna want to have a word with you about that haha. No seriously, I'm not sure why you keep insisting on that man. Ivy scales the same with RAM speed as previous Intel releases (and in my experience as AMD's too for that matter) It's a pretty linear thing. 1% perf boost for every 100Mhz in RAM speed. Doesn't matter if it's C2D, K10, Sandy Bridge or HAL9000

[Capt Lalle 0]

Ok thank you, I've ordered 2600 MHz Ram now and also hope to get a good CPU batch for Ivy overclocking.

[OmniAtlas 93]

I have 1333 ram now. What if I upgrade to 2600? How noticeable will it be?

[waygo 1]

There wont be any noticable benefit to fsx. That being said, faster ram does have it's benefits elsewere. I use it for video editing.

 

You're better off though with 2400 over 2600. I am prone to error, but this is what I learned when trying to figure out how to compare ram...

 

Take the CL/MHz*200. This allows you to truly compare different ram module speeds.

 

For example, the lowest cas latency for a 2600 on newegg is 10. 10/2600*200=.769 at $165 for 8gb

G.Skill Trident sells a 9CL 2400. 9/2400*200=.75 at $95 for 8gb

The 2400 would be faster AND cheaper.

 

Anyone can feel free to correct me if Im wrong.

 

hope to get a good CPU batch for Ivy overclocking.

 

What Mobo are you going to be putting it in and what sort of cooling did you have in mind. These two things will play a bigger difference in a good oc than what batch you get.

[Capt Lalle 0]

What Mobo are you going to be putting it in and what sort of cooling did you have in mind. These two things will play a bigger difference in a good oc than what batch you get.

I ordered the Asus Maximus V Gene and Corsair H100i watercooler. First I will look at the temps when overclocking without delidding, but I have this in mind if temps are not adequate with a stable 24/7 overclocking.

 

As for the Ram, they are already along the way, but thank you for your estimation. I will see how they go.

[waygo 1]

Thats pretty close to what I have. (Maximus Formula & h100i) I've been able to OC to 4.6 with no trouble. Asus has an auto tune feature that is supposed to auto OC to 4.6, but I found it wasn't stable during flights. I got better results on my own.

[TechguyMaxC 516]

I have 1333 ram now. What if I upgrade to 2600? How noticeable will it be?

 

Do you have an Ivy Bridge processor? If so, you can expect anywhere from 10-15% higher FPS in FSX according to the results obtained via FSXMark 11 testing (and recorded in that thread).

 

There wont be any noticable benefit to fsx.

 

This is incorrect. As mentioned above, FSXMark 11 testing has shown appreciable benefits to faster RAM, at least when paired with an Ivy Bridge processor (and a Z77 motherboard while we're at it).

 

You're better off though with 2400 over 2600. I am prone to error, but this is what I learned when trying to figure out how to compare ram...

 

Take the CL/MHz*200. This allows you to truly compare different ram module speeds.

 

For example, the lowest cas latency for a 2600 on newegg is 10. 10/2600*200=.769 at $165 for 8gb

G.Skill Trident sells a 9CL 2400. 9/2400*200=.75 at $95 for 8gb

The 2400 would be faster AND cheaper.

 

Anyone can feel free to correct me if Im wrong.

 

This is also incorrect. The CAS latency setting you're referencing is not a coefficient which can be applied to the clock speed in order to determine a generic "total speed" value, despite what some information on these and other forums may indicate. That being said, RAM with a lower CAS rating is preferable to that with a higher rating, given the same clock speed. There is truth in that there is a trade off to be made here, and I personally wouldn't take say 2600 RAM with CAS 13 vs. say 2400 RAM with CAS 9, but that's a much larger difference in latency than the example you indicated.

[MarkW 1,490]

Imagine if we had to run FSX on the HAL9000.

 

HAL9000 "I've lowered the landing flaps, Mark"

 

Me " HAL, for the love of god, we are above min flap extension speed, they will be torn off"

 

HAL9000 "I am opening the doors now, Mark."

 

Me "HAL, we are all going to die now there is no Oxygen at this altitude"

[waygo 1]

Do you have an Ivy Bridge processor? If so, you can expect anywhere from 10-15% higher FPS in FSX according to the results obtained via FSXMark 11 testing (and recorded in that thread).

 

 

 

This is incorrect. As mentioned above, FSXMark 11 testing has shown appreciable benefits to faster RAM, at least when paired with an Ivy Bridge processor (and a Z77 motherboard while we're at it).

 

 

 

This is also incorrect. The CAS latency setting you're referencing is not a coefficient which can be applied to the clock speed in order to determine a generic "total speed" value, despite what some information on these and other forums may indicate. That being said, RAM with a lower CAS rating is preferable to that with a higher rating, given the same clock speed. There is truth in that there is a trade off to be made here, and I personally wouldn't take say 2600 RAM with CAS 13 vs. say 2400 RAM with CAS 9, but that's a much larger difference in latency than the example you indicated.

 

The information I got about the ram came from this article at AnadTech: http://www.anandtech...igp-with-gskill

 

The idea is to find out which is truly faster, such as the example of cl9 2400 vs cl10 2600. The misconception is that the higher the MHz the fast the ram. But cl is an important, often overlook factor. According to the article reviewing ram for ivy bridge, the formula is: 2000*(cl/Mhz)=latencyin nanseconds. Why do you feel that is incorrect? (No disrespect intended, only trying to learn myself)

 

Also regarding the difference in fsx, I was responding to the original post asking about 2400 vs 2600, not the post referring to 1333 vs 2600 (that post popped up while I was typing and I did not see it until after I posted). I do not believe anyone will get a 10-15% increase in fps going from 2400-2600. Than again if you were only getting 10 fps switching from 2400-2600 got you up to 11, that would be 10% so I guess it could be possible. Likely they would not even notice the change though. Regarding 1333, that should absolutly be upgraded.

[TechguyMaxC 516]

Imagine if we had to run FSX on the HAL9000.

 

HAL9000 "I've lowered the landing flaps, Mark"

 

Me " HAL, for the love of god, we are above min flap extension speed, they will be torn off"

 

HAL9000 "I am opening the doors now, Mark."

 

Me "HAL, we are all going to die now there is no Oxygen at this altitude"

 

You should post this in the fly by wire discussion thread :lol:

 

The information I got about the ram came from this article at AnadTech: http://www.anandtech...igp-with-gskill

 

The idea is to find out which is truly faster, such as the example of cl9 2400 vs cl10 2600. The misconception is that the higher the MHz the fast the ram. But cl is an important, often overlook factor. According to the article reviewing ram for ivy bridge, the formula is: 2000*(cl/Mhz)=latencyin nanseconds. Why do you feel that is incorrect? (No disrespect intended, only trying to learn myself)

 

Also regarding the difference in fsx, I was responding to the original post asking about 2400 vs 2600, not the post referring to 1333 vs 2600 (that post popped up while I was typing and I did not see it until after I posted). I do not believe anyone will get a 10-15% increase in fps going from 2400-2600. Than again if you were only getting 10 fps switching from 2400-2600 got you up to 11, that would be 10% so I guess it could be possible. Likely they would not even notice the change though. Regarding 1333, that should absolutly be upgraded.

 

The speed increases I mentioned were in regard to switching from 1333 to a higher speed. I saw a 10% increase in FSXMark 11 average FPS moving from 1333 to 2133. I now run 2600, and saw a further 2-3% (or so) moving from 2133 to 2600.

 

As for the latency discussion, you have to remember that clock speed affects latency, so higher clocked RAM with higher listed latencies may not in fact have a higher latency in actual use. For example, 1333 CL 7 and 2600 CL 11 yield a much faster average latency on the 2600 RAM because those clock cycles are much shorter when you have for example 2600 (million) of them per second vs. 1333 (million) of them per second. Of course in this example the actual number of clock cycles is doubled due to the way DDR RAM is marketed and I have represented the RAM in the more common rating system for simplicity's sake.

[waygo 1]

As for the latency discussion, you have to remember that clock speed affects latency, so higher clocked RAM with higher listed latencies may not in fact have a higher latency in actual use.

 

Thank you for the explanation. What I don't understand though is how 2000*(cl/Mhz)=ns is wrong though. It allows you to generally compare two rams which advertise different speeds and latencies. That seems practical because until you buy them, what other way do you have of comparing apples to oranges. What you are describing if Im understanding you, is that the ram MAY be faster after you get it home in your machine and put it to use.

 

Right now I am using G.Skill Trident 2400 cl 9. Do you believe I would notice a diffence in fsx if I switched to 2600 cl 10?