Jump to content

Sign in to follow this  
FoxMulder1982

New PC for P3D - 9900k - RAM question

Recommended Posts

2 hours ago, FoxMulder1982 said:

🙂

as usual, I have additional question 🙂

I have read on some overclockers forum that if CPU and MOBO are only dual-chanel it is better to buy only 2 memory modules instead of 4 because it there is higher probability that 2 module setup will work better.

Based on your experience, can you confirm it?

 

I think the rule of thumb for best performance these days is to populate all four slots with a 4 stick kit, ie. 4x8GB (32gb) for example.  But you can buy and happily use a two stick kit.


    ROG Maximus X Apex Z370 -- 8086 @ 5.3 / NB 5.0 -- GSkill  @ 4133 c17-17-32~Cr1 1.42v  -- EVGA 1080Ti 6393 -- ROG PG279Q 1440P 150hz -- Corsair H100i V2 --Samsung EVO 850(s) -- Windows7 Pro 64 --Corsair 750X

Ken C

Share this post


Link to post
Share on other sites
3 hours ago, FunknNasty said:

I think the rule of thumb for best performance these days is to populate all four slots with a 4 stick kit, ie. 4x8GB (32gb) for example.  But you can buy and happily use a two stick kit.

That's exactly the opposite of both my experience and what I'd characterize as the long-standing conventional wisdom.

The higher the number of devices (DIMM modules), the more variability you introduce, the higher the current load on the IMC, and the more possibility that the timing ranges on all the devices won't have enough overlap to find a common set of timings where they all will run reliably (at least not anywhere near the rated overclocked speed).  I've never seen a manufacturer with a top-end 4-DIMM set that was as fast as the same manufacturer's 2-DIMM matched set using the same ICs.

Most manufacturers will tell you not to try and combine two 2-DIMM sets to make a 4-DIMM set...and if you do, don't expect it to run at its advertised (overclocked) settings.

Regards

  • Upvote 1

Bob Scott | AVSIM Forums Administrator | AVSIM Board of Directors

ATP Gulfstream II-III-IV-V

System: i9-10900K @ 5.2GHz on custom water loop, ASUS Maximus XII Hero, 32GB GSkill 3600MHz CAS15, eVGA 2080Ti XC Ultra, Samsung 55" JS8500 4K TV@30Hz, 5xSamsung SSD, eVGA 1KW PSU, 1Gbps internet

SB XFi Titanium, optical link to Yamaha RX-V467, Polk/Klipsch 6" bookshelf spkrs, Polk 12" subwoofer, 12.9" iPad Pro, PFC yoke/throttle quad/pedals with custom Hall sensors, Coolermaster HAF932 case, Stream Deck XL button box

Share this post


Link to post
Share on other sites
2 hours ago, w6kd said:

That's exactly the opposite of both my experience and what I'd characterize as the long-standing conventional wisdom.

The higher the number of devices (DIMM modules), the more variability you introduce, the higher the current load on the IMC, and the more possibility that the timing ranges on all the devices won't have enough overlap to find a common set of timings where they all will run reliably (at least not anywhere near the rated overclocked speed).  I've never seen a manufacturer with a top-end 4-DIMM set that was as fast as the same manufacturer's 2-DIMM matched set using the same ICs.

Most manufacturers will tell you not to try and combine two 2-DIMM sets to make a 4-DIMM set...and if you do, don't expect it to run at its advertised (overclocked) settings.

Regards

Not the pros at Asus. I believe the OP is/or going to use a Maximums XI Formula .....do yourself a favor and look it up over at the Rog forum.

......and I'll refrain from ever making a hardware recommendation on these forums again.


    ROG Maximus X Apex Z370 -- 8086 @ 5.3 / NB 5.0 -- GSkill  @ 4133 c17-17-32~Cr1 1.42v  -- EVGA 1080Ti 6393 -- ROG PG279Q 1440P 150hz -- Corsair H100i V2 --Samsung EVO 850(s) -- Windows7 Pro 64 --Corsair 750X

Ken C

Share this post


Link to post
Share on other sites

Think i have this up before, its pefectly fine to do your own matching mem modules.

example, its one slot that read the xmp take one stick 3200 c14 and second 3600 c15 start the pc enable xmp it read 3200 c14 then shut down and swap the place of the mems start pc enable xmp it read 3600 c15.

give you an example on my X299 system i have 32gb 4X8gb 2sticks 3600 c15 and 2sticks 3600 c16 tested them and burn two new xmp tables on all 4 sticks one 4000 c16 and second  4133 c16 ( for Z370-390).

binning mems how its done, buy lot of differnt samsung b die for examle 10 kits 3200 10 kits 3600 3kits 4266 then test the modules on by one pick the best 4 sticks then burn yor own xmp ex 4133 c15 this set can be two 3200 one 3600 one 4266.

In short term if you have samsung B-die with the same type of pcb you can place them whatever you want the only thing that matters is the quality of the weakest chip on the memory module.

 

The Latency calc cl x 2000 / ramspeed is only tumb of rule not really accurate we all now that Comand rate 1 vs 2 do a big difference and its more timings that counts sub end tetra for example and even the latency in the IMC

Edited by westman

Share this post


Link to post
Share on other sites
8 hours ago, fghdgdfdfgfgf said:

......and I'll refrain from ever making a hardware recommendation on these forums again.

Why?  Just because someone might respectfully present experiences or opinions that differ from yours?

The last thing we need here is to become an echo chamber.

Regards


Bob Scott | AVSIM Forums Administrator | AVSIM Board of Directors

ATP Gulfstream II-III-IV-V

System: i9-10900K @ 5.2GHz on custom water loop, ASUS Maximus XII Hero, 32GB GSkill 3600MHz CAS15, eVGA 2080Ti XC Ultra, Samsung 55" JS8500 4K TV@30Hz, 5xSamsung SSD, eVGA 1KW PSU, 1Gbps internet

SB XFi Titanium, optical link to Yamaha RX-V467, Polk/Klipsch 6" bookshelf spkrs, Polk 12" subwoofer, 12.9" iPad Pro, PFC yoke/throttle quad/pedals with custom Hall sensors, Coolermaster HAF932 case, Stream Deck XL button box

Share this post


Link to post
Share on other sites
9 hours ago, westman said:

binning mems how its done, buy lot of differnt samsung b die for examle 10 kits 3200 10 kits 3600 3kits 4266 then test the modules on by one pick the best 4 sticks then burn yor own xmp ex 4133 c15 this set can be two 3200 one 3600 one 4266.

In short term if you have samsung B-die with the same type of pcb you can place them whatever you want the only thing that matters is the quality of the weakest chip on the memory module.

The Latency calc cl x 2000 / ramspeed is only tumb of rule not really accurate we all now that Comand rate 1 vs 2 do a big difference and its more timings that counts sub end tetra for example and even the latency in the IMC

I'm certain that if one has the ability to sort through a workbench full of motherboards, CPUs, and memory DIMMs, that a lot of magic can be worked by binning and matching and tweaking the best of each.  But for the average system builder that will order parts from an online supplier or buy them at a retail store without the resources/ability to pick and choose, I think it's still better to improve your odds of a good memory overclock by sticking to a factory-matched two-DIMM solution with true latency as a metric for expected performance.

Sure, the true latency approximation is a rule of thumb, but it's a good one for the average builder.  IMC latency will be the same no matter what DIMM you choose if your target system for either is the same mobo+CPU, command rate rarely comes into play with memory overclocks at high memory clock frequencies, and most people are not going to mess with sub-timings at all due to the time required and the significant in-depth knowledge needed to tweak them.

If one is not a competitive benchmarker or a professional system-builder, then a lot of the techniques that you describe to build a cutting-edge system would be out of reach.  The average Joe (or Jill) just wants to get their sim running as well as possible within the range of variability of parts procured through a normal retail sale.

Cheers

 


Bob Scott | AVSIM Forums Administrator | AVSIM Board of Directors

ATP Gulfstream II-III-IV-V

System: i9-10900K @ 5.2GHz on custom water loop, ASUS Maximus XII Hero, 32GB GSkill 3600MHz CAS15, eVGA 2080Ti XC Ultra, Samsung 55" JS8500 4K TV@30Hz, 5xSamsung SSD, eVGA 1KW PSU, 1Gbps internet

SB XFi Titanium, optical link to Yamaha RX-V467, Polk/Klipsch 6" bookshelf spkrs, Polk 12" subwoofer, 12.9" iPad Pro, PFC yoke/throttle quad/pedals with custom Hall sensors, Coolermaster HAF932 case, Stream Deck XL button box

Share this post


Link to post
Share on other sites
19 minutes ago, w6kd said:

I'm certain that if one has the ability to sort through a workbench full of motherboards, CPUs, and memory DIMMs, that a lot of magic can be worked by binning and matching and tweaking the best of each.  But for the average system builder that will order parts from an online supplier or buy them at a retail store without the resources/ability to pick and choose, I think it's still better to improve your odds of a good memory overclock by sticking to a factory-matched two-DIMM solution with true latency as a metric for expected performance.

Sure, the true latency approximation is a rule of thumb, but it's a good one for the average builder.  IMC latency will be the same no matter what DIMM you choose if your target system for either is the same mobo+CPU, command rate rarely comes into play with memory overclocks at high memory clock frequencies, and most people are not going to mess with sub-timings at all due to the time required and the significant in-depth knowledge needed to tweak them.

If one is not a competitive benchmarker or a professional system-builder, then a lot of the techniques that you describe to build a cutting-edge system would be out of reach.  The average Joe (or Jill) just wants to get their sim running as well as possible within the range of variability of parts procured through a normal retail sale.

Cheers

 

Sorry, Bob  i agree with Ken in this , i only tried to explain how it works , but the moderator here seems bash out things that was from the past.

your own words: That's exactly the opposite of both my experience and what I'd characterize as the long-standing conventional wisdom.

Thanks

 

Share this post


Link to post
Share on other sites
37 minutes ago, westman said:

Sorry, Bob  i agree with Ken in this , i only tried to explain how it works , but the moderator here seems bash out things that was from the past.

Sorry, but this is not making sense.  No idea who "Ken" is.  And there has been no moderation activity in this thread.  None.

Edit:  OK, Ken, aka "fghdgdfdfgfgf" aka "FunknNasty"


Bob Scott | AVSIM Forums Administrator | AVSIM Board of Directors

ATP Gulfstream II-III-IV-V

System: i9-10900K @ 5.2GHz on custom water loop, ASUS Maximus XII Hero, 32GB GSkill 3600MHz CAS15, eVGA 2080Ti XC Ultra, Samsung 55" JS8500 4K TV@30Hz, 5xSamsung SSD, eVGA 1KW PSU, 1Gbps internet

SB XFi Titanium, optical link to Yamaha RX-V467, Polk/Klipsch 6" bookshelf spkrs, Polk 12" subwoofer, 12.9" iPad Pro, PFC yoke/throttle quad/pedals with custom Hall sensors, Coolermaster HAF932 case, Stream Deck XL button box

Share this post


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
Sign in to follow this  

  • Tom Allensworth,
    Founder of AVSIM Online


  • Flight Simulation's Premier Resource!

    AVSIM is a free service to the flight simulation community. AVSIM is staffed completely by volunteers and all funds donated to AVSIM go directly back to supporting the community. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. Thank you for your support!

    Click here for more information and to see all donations year to date.
  • Donation Goals

    AVSIM's 2020 Fundraising Goal

    Donate to our annual general fundraising goal. This donation keeps our doors open and providing you service 24 x 7 x 365. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. We reset this goal every new year for the following year's goal.


    27%
    $6,985.00 of $25,000.00 Donate Now
×
×
  • Create New...