[RESOLVED] Ram Timings question so be ready.
Results 1 to 8 of 8

Thread: [RESOLVED] Ram Timings question so be ready.

  1. #1
    Join Date
    Nov 2018
    Location
    Chicago, Il.
    Posts
    256

    Resolved [RESOLVED] Ram Timings question so be ready.

    Finally a topic I do know a lot about but still have questions.

    I had bought a set of matched sticks of ram 4 sticks total. (or so I was told. I believe they swapped the heat spreaders on 2 of them). They were all supposed to be GSkill RipjawsX Number 1600C9-8GXM. It turns out that 2 of them are 12800CL10-8GBXL. These 2 sets do not have matching timings and are not running at the same timings. On the 1600C9's and 12800CL10's here are the differences side by side running and not matching.
    @1600
    CH A CH B

    1600C9 | 12800CL10
    CL9 | CL10
    RCD9 | RCD10
    RP9 | RP10
    RAS9 | RAS30
    ~~~~~~~~~~~~~

    They are running in 2 different channels this way. From there the RC, RFC,RRD etc gets even worse off. Now I can force feed these to match at 1333 (667). Would this be the better option? I really don't feel like pulling this hulk of a HSF off of this to swap ram out.

    Any input would help.

  2. #2
    Join Date
    Feb 2000
    Location
    Idaho Falls, Idaho, USA
    Posts
    18,063
    This might depend on your ultimate goal. Is the system running well and at a decent speed as it is now? If yes, I would probably leave things as they are. If you have lots of spare time and don't mind running extensive memory testing afterwards, you could try to get the timings to match up, and then test to see what happens. Note that without extensive testing I would not just start using the system for production work, since any small memory error might corrupt a document or database file, etc.

  3. #3
    Join Date
    Nov 2018
    Location
    Chicago, Il.
    Posts
    256
    It runs yes, but it runs sloppy. In games it stutters when it shouldn't (occlusion, shadows etc off), hiccups sometimes on heavy web-based applications, just things that it shouldn't be doing. I did go ahead and change some of the timings already and am testing at 1333/667. One of the timings confused me however. The tRFC from AIDA tells me that it runs around 174 which seems insanely high for a read time, I could see 100 (even though that still seems high. For 1600 it should be, as I lower the speed of the ram the faster the read or write times should be or so I thought). So yeah I am having an issue here. Also when I enter the bios the entry for the tRFC is tRFC0 and tRFC1 and both of them are set to 4. Not to sure on what to believe here.

  4. #4
    Join Date
    Nov 2018
    Location
    Chicago, Il.
    Posts
    256
    The RAS in the original post for the 1600C9 was supposed to be 24. I just noticed that.

  5. #5
    Join Date
    Nov 2018
    Location
    Chicago, Il.
    Posts
    256
    Meanwhile back at the lab,
    After extensive testing, reboots, timing and speed adjustments later.....

    I have found the place for the ram finally until I do pull the HSF off. The XMP timings were just not going to cut it. After thorough research, I did find that I could program the speeds I needed but with great risk. It does run 100% better at 1333 with some minor tweaks.

    So...until I get off my butt and pull this thing apart again I will just limp for now.

  6. #6
    Join Date
    Nov 2018
    Location
    Chicago, Il.
    Posts
    256
    Ok, so I do have a question now. I am going to pull it apart in the morning because I am going to get this board out of here. It's on a 4+1 phase so I am moving to an 8+2 Phase board for obvious reasons. Doesn't it make sense to just run the 16 Gb (2x8Gb 9-9-9-24 @ 1600) rather than all 4 at 1333 and having the timing issues and having to tweak it again? I don't know if the board will handle it differently and do plan on benching it first before it is installed into the case. Is it even worth testing?

  7. #7
    Join Date
    Mar 2009
    Location
    Arkham Asylum, Cell 13
    Posts
    11,686
    Assuming you can't return the RAM for the correct model, I would run all 32GB at the 1333 timings. It depends on what you are trying to do, but I would rather have more RAM total.

    The only way to know the difference for sure is to benchmark your computer with the 1600/16GB and 1333/32GB.

  8. #8
    Join Date
    Nov 2018
    Location
    Chicago, Il.
    Posts
    256
    Yep. It stayed at 1333. Tested for 2 weeks with various ram settings. 1333 it is for now.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •