If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
too-fast RAM will not work?
I received a feeble PC - AMD E2 CPU with a jogging 1866 speed memory
controller. Hoping that I could improve it somehow, I stuck in a second DDR4 DIMM. The spare I had was 2666 speed, but didn't work. My understanding is that DIMMs offer 2 speeds below the rated speed, so it would have offered 2133 and 2400. I went to local PC shop and bought the cheapest 4 GB 2400 DIMM they had, and that worked. |
#2
|
|||
|
|||
too-fast RAM will not work?
wrote:
I received a feeble PC - AMD E2 CPU with a jogging 1866 speed memory controller. Hoping that I could improve it somehow, I stuck in a second DDR4 DIMM. The spare I had was 2666 speed, but didn't work. My understanding is that DIMMs offer 2 speeds below the rated speed, so it would have offered 2133 and 2400. I went to local PC shop and bought the cheapest 4 GB 2400 DIMM they had, and that worked. http://www.cpu-world.com/Cores/Stoney_Ridge.html "Stoney Ridge processors have only 2 CPU cores, 1 MB L2 cache, up to 3 GPU compute units (192 shaders), and support single-channel DDR4 memory with maximum data rate of 2133 MHz. " The two DIMMs sit on a common channel, and the BIOS has to select settings below the normal value for either DIMM by itself. It can relax Tcas or Trcd or even attempt to drop the clock rate below normal. This is an AMD tradition (dropping clock when two DIMMs sit on a channel). In DDR400 days, it was typical to drop to DDR333 setting when two DIMMs were present on a channel. You can drop the clock a lot, before the PLL packs it in. The BIOS can compute acceptable settings, other than the data points in the SPD table. For example, it's possible a DDR400 would run at DDR160 or so, just to give some idea how far it could stretch. (You can find this information in a Micron or Samsung DRAM chip datasheet.) If the design uses a DLL for controlling phase, if you drop the frequency too much, it runs out of taps to select. So there is a method to their madness. There are hardware resource limits which are generous, but not infinite. For kicks, test the 2666 DIMM by itself and see if the BIOS can configure it. There's a remote possibility the 2666 will come up by itself. And that's how I would have carried out my first test case, is the 2666 by itself. You can always dump the SPD tables and admire the parameters yourself, and seek an answer to why it did that. You'll need a "JEDEC decoder" PDF to help you. Some JEDEC info is members only. Some info is "public" but you need an account to do a download (which is intended to limit their bandwidth exposure, if they opened it up for fully public download). It's really a shame they couldn't just move the useful stuff to a separate server, so the handful of decoders would be there for public consumption. Paul |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
A fast trick that did not work | Lars[_2_] | Storage (alternative) | 12 | July 19th 09 10:36 AM |
Why does hd defrag work when they work on virtual/logical blocks rather than physical? | Anthony Paul | Storage (alternative) | 17 | December 27th 06 10:28 PM |
ATI All In Wonder 8500DV problems: I cannot get this card to work in Windows2000,except if I turn the acceleration all the way off.I have tried everything. It will work for about 10-20 minutes,and then all the text just goes berzerk.Looks kinda like | TheWalrus | Ati Videocards | 3 | November 8th 04 07:55 PM |
I can't get my dazzle card (FAST Dazzle dv.now-AV V1.4) to work? | Alistair Smythe | Homebuilt PC's | 1 | June 18th 04 12:19 PM |
FDISK /MBR and fixmbr did not work, but I have a little plan that might work. | Harry Nikken | General Hardware | 2 | December 23rd 03 11:28 AM |