Jump to content

Ram Timings

Hi,

Whats the difference between CL 15-17-17 and CL 15-16-16 I assume the second one is better?

 

 

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

15 minutes ago, PeggersXtreme said:

Hi,

Whats the difference between CL 15-17-17 and CL 15-16-16 I assume the second one is better?

 

 

I think I get it each number is the amount of ns it takes to complete an operation

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

Technically the second one is slightly faster. You should not mix memory modules with different timings. (Generally they will all be slowed down to the slowest one in your lot, some memory sticks don't like to change timings.)

 

Post next post edit: Assuming all else being equal the above will generally be correct.

Link to comment
Share on other sites

Link to post
Share on other sites

9 minutes ago, corrado33 said:

Technically the second one is slightly faster. You should not mix memory modules with different timings. (Generally they will all be slowed down to the slowest one in your lot, some memory sticks don't like to change timings.)

Technically, the second one isn't faster. You don't know the raw bandwidth of the kit, nor can you possibly know the tertiary timings (or the board training for that matter). Primary timings by themselves mean very little. I can completely loosen my primary timings, and focus my effort entirely in tertiary timings, and not only smash people in terms of bandwidth, but even my latency would be far lower (even with a higher cas latency).

 

Even if both kits have the exact same speeds (Let's use DDR4 3000mhz as an example), different IC's act differently when it comes to tertiary timings. You might see Samsung kits that like their tertiary timings very low, while some SK Hynix IC's perform worse when going too low. 

 

All of this is ignoring the gigantic elephant in the room, RTL/IO-L offsets, though your board and IMC will be the biggest factor in that regard. My point is, primary timings only make up a very tiny fraction of overall memory speed. You get far more out of your ram from the tertiary timings, which are never advertised (mostly because it's impossible to advertise them due to their training being dictated by IMC and board). 

My (incomplete) memory overclocking guide: 

 

Does memory speed impact gaming performance? Click here to find out!

On 1/2/2017 at 9:32 PM, MageTank said:

Sometimes, we all need a little inspiration.

 

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

21 minutes ago, MageTank said:

Technically, the second one isn't faster. You don't know the raw bandwidth of the kit, nor can you possibly know the tertiary timings (or the board training for that matter). Primary timings by themselves mean very little. I can completely loosen my primary timings, and focus my effort entirely in tertiary timings, and not only smash people in terms of bandwidth, but even my latency would be far lower (even with a higher cas latency).

 

Even if both kits have the exact same speeds (Let's use DDR4 3000mhz as an example), different IC's act differently when it comes to tertiary timings. You might see Samsung kits that like their tertiary timings very low, while some SK Hynix IC's perform worse when going too low. 

 

All of this is ignoring the gigantic elephant in the room, RTL/IO-L offsets, though your board and IMC will be the biggest factor in that regard. My point is, primary timings only make up a very tiny fraction of overall memory speed. You get far more out of your ram from the tertiary timings, which are never advertised (mostly because it's impossible to advertise them due to their training being dictated by IMC and board). 

Ok. I had assumed "all else being equal." Should have written that down. ;) 

Link to comment
Share on other sites

Link to post
Share on other sites

On ‎12‎/‎12‎/‎2016 at 9:29 PM, corrado33 said:

Technically the second one is slightly faster. You should not mix memory modules with different timings. (Generally they will all be slowed down to the slowest one in your lot, some memory sticks don't like to change timings.)

 

Post next post edit: Assuming all else being equal the above will generally be correct.

I haven't mixed rams lol I was looking to buy some and came across those timings

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

On ‎12‎/‎12‎/‎2016 at 9:44 PM, MageTank said:

Technically, the second one isn't faster. You don't know the raw bandwidth of the kit, nor can you possibly know the tertiary timings (or the board training for that matter). Primary timings by themselves mean very little. I can completely loosen my primary timings, and focus my effort entirely in tertiary timings, and not only smash people in terms of bandwidth, but even my latency would be far lower (even with a higher cas latency).

 

Even if both kits have the exact same speeds (Let's use DDR4 3000mhz as an example), different IC's act differently when it comes to tertiary timings. You might see Samsung kits that like their tertiary timings very low, while some SK Hynix IC's perform worse when going too low. 

 

All of this is ignoring the gigantic elephant in the room, RTL/IO-L offsets, though your board and IMC will be the biggest factor in that regard. My point is, primary timings only make up a very tiny fraction of overall memory speed. You get far more out of your ram from the tertiary timings, which are never advertised (mostly because it's impossible to advertise them due to their training being dictated by IMC and board). 

So are you saying a CL16 kit could actually be faster than a CL15 kit? I wanted crucial ballistix elite but was told the CL16 timings were terrible and that I should get the corsair vengeance at CL15 problem is I can't run rams at 3000mhz on my x99 platform it is only stable at 2666mhz so the corsair kit would run at 2666mhz the same as the crucial kit

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

4 hours ago, PeggersXtreme said:

So are you saying a CL16 kit could actually be faster than a CL15 kit? I wanted crucial ballistix elite but was told the CL16 timings were terrible and that I should get the corsair vengeance at CL15 problem is I can't run rams at 3000mhz on my x99 platform it is only stable at 2666mhz so the corsair kit would run at 2666mhz the same as the crucial kit

What I meant by that, is CL is not the be all, end all measurement of total latency. You can have higher CL, but lower latency by increasing the raw clock speed, or tightening tertiary timings/RTL/IO-L offsets. Crucial uses Micron IC's, and are often bad at manual overclocking, which is why most people recommend against them. That being said, if they are 1.2v, they have overclocking headroom in them. Simply change voltage to 1.35v, and tinker with the timings. I have a spare set of Crucial Ballistix myself that went from 2133 C15 to 3000 C16-18-18-36. Now, as you stated, X99 will have a difficult time pulling this off with it's IMC, unless you use Broadwell-E with a board that can handle higher speeds. Haswell-E's IMC is simply incapable of driving high speed memory without making sacrifices in bandwidth elsewhere (again, tertiary timings).

 

Another thing to point out, is that Corsair also uses Micron IC's in their cheaper kits. It's a gamble whether or not you will get Micron or Hynix in a cheap Corsair kit, so for what it's worth, both kits might have the exact same IC's. There's not a doubt in my mind that if you buy the Crucial kit, slightly increase volts, and slightly tighten the primaries, that you could get it to run exactly the same as that Corsair kit, while having the aesthetics of the Crucial kit. 

My (incomplete) memory overclocking guide: 

 

Does memory speed impact gaming performance? Click here to find out!

On 1/2/2017 at 9:32 PM, MageTank said:

Sometimes, we all need a little inspiration.

 

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, MageTank said:

What I meant by that, is CL is not the be all, end all measurement of total latency. You can have higher CL, but lower latency by increasing the raw clock speed, or tightening tertiary timings/RTL/IO-L offsets. Crucial uses Micron IC's, and are often bad at manual overclocking, which is why most people recommend against them. That being said, if they are 1.2v, they have overclocking headroom in them. Simply change voltage to 1.35v, and tinker with the timings. I have a spare set of Crucial Ballistix myself that went from 2133 C15 to 3000 C16-18-18-36. Now, as you stated, X99 will have a difficult time pulling this off with it's IMC, unless you use Broadwell-E with a board that can handle higher speeds. Haswell-E's IMC is simply incapable of driving high speed memory without making sacrifices in bandwidth elsewhere (again, tertiary timings).

 

Another thing to point out, is that Corsair also uses Micron IC's in their cheaper kits. It's a gamble whether or not you will get Micron or Hynix in a cheap Corsair kit, so for what it's worth, both kits might have the exact same IC's. There's not a doubt in my mind that if you buy the Crucial kit, slightly increase volts, and slightly tighten the primaries, that you could get it to run exactly the same as that Corsair kit, while having the aesthetics of the Crucial kit. 

is Kingston hyper x savage or Gskill V series any better I plan to move away from x99 onto z170 or z270 I would've waited for zen but after seeing all the benchmarks against 6900k it looks like zen will be a weak single core chip but good for multithreading just my take from what I've seen anyway.

:Edit:  Also Gskil Trident z white and grey kit at CL14 3200mhz

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, PeggersXtreme said:

is Kingston hyper x savage or Gskill V series any better I plan to move away from x99 onto z170 or z270 I would've waited for zen but after seeing all the benchmarks against 6900k it looks like zen will be a weak single core chip but good for multithreading just my take from what I've seen anyway.

:Edit:  Also Gskil Trident z white and grey kit at CL14 3200mhz

G Skill Ripjaws V 3200 C14 is what I bought. In fact, those kits have even more overclocking headroom in them. I increased voltage from 1.35 to 1.39, and took my kit from 3200 C14, to 3600 C14-14-14-28-2. Combine these tight primaries with tighter tertiary timings, and you end up with bandwidth like this:

 

r2eh9qc.jpg

 

Samsung B-Die (what those kits use) overclocks quite well, with very little effort. Sadly, on X99, you won't get much out of a kit like that. Once you do upgrade to Z170/Z270 though, you will most definitely be able to achieve 3200 and beyond. As for Kingston Hyper X Savage, I believe those use Hynix AFR/MFR. As for overclocking, they do great, far better than Micron, but they might fall behind Samsung B-Die on air. I can't remember, and I am sure someone will correct me if I am wrong. Best place to find out information like that, would be under the memory subsection of the HWbot forums. 

My (incomplete) memory overclocking guide: 

 

Does memory speed impact gaming performance? Click here to find out!

On 1/2/2017 at 9:32 PM, MageTank said:

Sometimes, we all need a little inspiration.

 

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

42 minutes ago, MageTank said:

G Skill Ripjaws V 3200 C14 is what I bought. In fact, those kits have even more overclocking headroom in them. I increased voltage from 1.35 to 1.39, and took my kit from 3200 C14, to 3600 C14-14-14-28-2. Combine these tight primaries with tighter tertiary timings, and you end up with bandwidth like this:

 

r2eh9qc.jpg

 

Samsung B-Die (what those kits use) overclocks quite well, with very little effort. Sadly, on X99, you won't get much out of a kit like that. Once you do upgrade to Z170/Z270 though, you will most definitely be able to achieve 3200 and beyond. As for Kingston Hyper X Savage, I believe those use Hynix AFR/MFR. As for overclocking, they do great, far better than Micron, but they might fall behind Samsung B-Die on air. I can't remember, and I am sure someone will correct me if I am wrong. Best place to find out information like that, would be under the memory subsection of the HWbot forums. 

Yeah I want to get a kit that I can move over to the new platform even if I have to run them at 2133 or 2400 to start with :D just don't know if I should go 6700k or hold out for the 7700k lol loving the look of the Asus Strix z270 boards though although I do prefer msi boards normally. Is it much work to get a 6700k to 4.5ghz the 7700k is apparently equal ipc clock for clock vs 6700k so I assume bringing them to same clock speeds would yield same performance.

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

36 minutes ago, PeggersXtreme said:

Yeah I want to get a kit that I can move over to the new platform even if I have to run them at 2133 or 2400 to start with :D just don't know if I should go 6700k or hold out for the 7700k lol loving the look of the Asus Strix z270 boards though although I do prefer msi boards normally. Is it much work to get a 6700k to 4.5ghz the 7700k is apparently equal ipc clock for clock vs 6700k so I assume bringing them to same clock speeds would yield same performance.

Yeah, 4.5ghz is very common. 4.6/4.7ghz is the average for 6700k's. As far as IPC goes, Skylake and Kaby are 100% the same in that regard. Zero difference. The number one thing Kaby has over Skylake, is the advance turbo boost feature which is far more efficient. For manual overclockers, this means absolutely nothing. For mobile SKU's, it's absolutely amazing. Means better performance and better battery life. 

 

I hit 4.5ghz in an ITX case, with a 45mm ITX cooler. Granted, my chip is delidded, you should be perfectly capable of hitting 4.5ghz in a larger case, with a decent air cooler. Nothing is ever guaranteed with overclocking, but the chances are pretty high. http://www.overclock.net/t/1570313/skylake-overclocking-guide-with-statistics

 

As you can see, average OC on Skylake is 4.68ghz, while median is 4.7ghz. Either way, good luck in advance with your new platform. Hope you enjoy it!

My (incomplete) memory overclocking guide: 

 

Does memory speed impact gaming performance? Click here to find out!

On 1/2/2017 at 9:32 PM, MageTank said:

Sometimes, we all need a little inspiration.

 

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

17 minutes ago, MageTank said:

Yeah, 4.5ghz is very common. 4.6/4.7ghz is the average for 6700k's. As far as IPC goes, Skylake and Kaby are 100% the same in that regard. Zero difference. The number one thing Kaby has over Skylake, is the advance turbo boost feature which is far more efficient. For manual overclockers, this means absolutely nothing. For mobile SKU's, it's absolutely amazing. Means better performance and better battery life. 

 

I hit 4.5ghz in an ITX case, with a 45mm ITX cooler. Granted, my chip is delidded, you should be perfectly capable of hitting 4.5ghz in a larger case, with a decent air cooler. Nothing is ever guaranteed with overclocking, but the chances are pretty high. http://www.overclock.net/t/1570313/skylake-overclocking-guide-with-statistics

 

As you can see, average OC on Skylake is 4.68ghz, while median is 4.7ghz. Either way, good luck in advance with your new platform. Hope you enjoy it!

If it gives me less hassle than my current platform I will be happy lol xD

 

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, PeggersXtreme said:

If it gives me less hassle than my current platform I will be happy lol xD

 

Well, if you ever have any questions about overclocking the ram, or any difficulty getting XMP to load, let me know. Memory seems to be one of the few things I understand when it comes to overclocking. 

My (incomplete) memory overclocking guide: 

 

Does memory speed impact gaming performance? Click here to find out!

On 1/2/2017 at 9:32 PM, MageTank said:

Sometimes, we all need a little inspiration.

 

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, MageTank said:

Well, if you ever have any questions about overclocking the ram, or any difficulty getting XMP to load, let me know. Memory seems to be one of the few things I understand when it comes to overclocking. 

ok thanks :D

Custom PC:

Intel Core i7 5960X GTX1080 Strix | 64gb 2400mhz DDR4 Corsair Vengeance Ram | Samsung S951 256gb M.2 | 4TB WD Blue SSHD | EVGA Supernova G2 1300watt PSU | Acer XB270HU | Corsair K95 RGB | Corsair M65 RGB |

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×