I'm hardware intermediate, please keep the language simple. This is a workstation for high res photos (sometimes over 10gb) and video. It is currently not connected to the net because every time windows updates my wireless goes out. That's for a different post. First is there a good free program I can benchmark the memory and test reliability? At least a place in server 2012 to see ram clockspeed?
dual quad core Xeon E5440 @ 2.83ghz not hyper threaded (should it be? Don't know what this does)
4x4gb dual channel fb-ddr2 PC2-5300 ECC DIMM QUAD RANK 667mhz
4x1gb ????? pc23-5300F 555-12-a1 not sure how to read that if it is 555mhz?
Windows Server 2012 Datacenter
First, I can't find anything in windows showing clock speed on the RAM so I used speccy and it is reading 332.5mhz. I started today with 8x1gb then removed the bottom 4 on the board and replaced with 4x4 dual channel fb-ddr2 PC2-5300 ECC DIMM QUAD RANK 667mhz. Speccy continues reading 332.5mhz. I removed the final 4x1 leaving on the new 4x4 blocked together in the bottom half and the top half empty.
I called the place I got the sticks from and they say being DDR they think it is reading actual clockspeed before adding the second DDR stream which doubles it so the 332.5 is technically correct and everything is fine. He says this board will downclock 800mhz and not even bother to boot if it really were 333.
I asked about the slot colors. He said without knowing all the specifics of the board, in general I should fill all of one color first then the other which would sit on the motherboard 4 1 4 1 4 1 4 1 to achieve . Someone replied on a post on here a long time ago to put Blues first, then blacks. Server Board S5000XVN/XSL. I don't understand why. To me, If they only need to be paired to run dual channel, a block of 4x4 are all matching blues to blacks anyway and it is moot. I also ran Belarc advisor and it shows a1 a2 b1 b2 occupied, c1 c2 d1 d2 empty, doesn't that prove it? Should I REALLY pull the ram out and go 4 1 4 1 4 1 4 1? The guy in that other post even made it sound like it would be faster if I left the 4x1gb out completely.
I was reading that SO-DIMMs are more or less equal in power and voltage rating to DIMMs, and as memory technology moves forward and despite their smaller size, SO-DIMM technology allows for performance speeds as high as the those of the larger DIMMs[clarification needed] (clock speeds such as 400 MHz PC-3200, for example, and CAS latency such as 2.0, 2.5, and 3.0) and capacity (512MB, 1GB, etc.).
The primary benefit of DDR3 SDRAM over its immediate predecessor, DDR2 SDRAM, is its ability to transfer data at twice the rate (eight times the speed of its internal memory arrays), enabling higher bandwidth or peak data rates. With two transfers per cycle of a quadrupled clock signal, a 64-bit wide DDR3 module may achieve a transfer rate of up to 64 times the memory clock speed megahertz (MHz) in megabytes per second (MB/s). With data being transferred 64 bits at a time per memory module, DDR3 SDRAM gives a transfer rate of (memory clock rate) × 4 (for bus clock multiplier) × 2 (for data rate) × 64 (number of bits transferred) / 8 (number of bits/byte). Thus with a memory clock frequency of 100 MHz, DDR3 SDRAM gives a maximum transfer rate of 6400 MB/s. In addition, the DDR3 standard permits DRAM chip capacities of up to 8 Gbit.
So in raw transfer speeds between RAM & CPU what did I do today? What is the before (8x1) and after speed (both 4x4DC 667 alone & 4x4+4x1)? What about strictly 32gb of the new?
Thanks again, Chris