Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Author Topic: RAM Latency comparisons in 24GB config VS 8GB and swapped channels  (Read 2650 times)

pelskein

  • Just arrived
  • *
  • Posts: 1
    • https://walmartone.me/

i lately went from a 8GB 2x4GB ( CMK8GX4M2B3000C15R ) config to a 24GB 2x4GB + an extra 2x8GB ( CMK16GX4M2B3000C15 ) config.

I hesitated at the start because of raised issues about multiplied latency as well as different issues with different sizes mixes and now not exactly matched memory kits.

besides, I used AIDA64 to degree memory latency, earlier than and after the improve. I additionally in comparison what occurs when you switch the kits channels at the mobo.

earlier than the improve, I measured the latency in AIDA64 ( now not through the gear > cache and memory benchmark ), but at once with the Menu > Benchmark > memory Latency benchmark.

I then measured it after the improve by way of setting the brand new 16GB package within the desired channel ( the one farthest from the cpu socket, A2/B2 in line with my mobo's manual, mobo being the MSI Z370 pc pro ), with the 8GB package in the different channel and additionally by swapping the channels so the 8GB kit is inside the preferred channel and the 16GB in the different channel.

The assessments have been performed at 2933mhz, the unique channels timings have been set to "unlink" in order that every channel had it is own SPD prescribed default timings.

The timings for the 8gb were: 1T/15/17/17/35/390 + secondary and tertiary timings on car ( stayed regular across all assessments ).

The timings for the 16gb had been: 1T/15/17/17/35/525 + secondary and tertiary timings on vehicle ( stayed steady across all tests ).

yes the 16GB is slower on default settings, no I couldn't get it all the way down to 390, it become introducing mistakes.

every test turned into ran three instances with the outcomes averaged, energy settings had been set to excessive performance, all viable heritage procedures had been killed, antivirus became off and disconnected from net to limit disturbances.

consequences:

8GB config - 52.3ns

24GB config, 16GB preferred - 59.2ns

24GB config, 8GB desired - 52.5ns

My speculation changed into that inside the 24GB, 8GB desired channel config, the faster 8GB kit gets used first. opposite for the 16GB desired channel config with the 16GB package crammed first and the quicker 8GB used last. plainly hypothesis become confirmed.

I concept this facts might prove useful to all and sundry who become considering upgrading from 8GB to 24GB or 4GB to 12GB or even 16GB to 48GB.

for max overall performance, vicinity the quickest lowest latency RAM within the mobo's favored channel slots as indicated in your mobo's guide.
Logged

Alex Atkin UK

  • Addicted Kitizen
  • *****
  • Posts: 5260
    • Thinkbroadband Quality Monitors
Re: RAM Latency comparisons in 24GB config VS 8GB and swapped channels
« Reply #1 on: May 03, 2020, 10:06:11 PM »

Interesting, although its a bad configuration either way.  I don't think there is any real-world benefit to testing latencies is a mismatched dual and single channel configuration like that, other than sheer curiosity.

I saw a pretty noticable improvement when upgrading my PCs from 2400Mhz RAM to 3200Mhz, especially so in my server I did today that has been running folding at home, a very CPU/memory intensive process.  It seems to be getting ~25% more performance out of the CPU.

Its not something you will always notice, but if you game it can dramatically improve frame drops/stutter.
Logged
Broadband: Zen Full Fibre 900 + Three 5G Routers: pfSense (Intel N100) + Huawei CPE Pro 2 H122-373 WiFi: Zyxel NWA210AX
Switches: Netgear MS510TXUP, Netgear MS510TXPP, Netgear GS110EMX My Broadband History & Ping Monitors