Intel Xeon E5-2670 Deal and Price Tracking
- Thread starter Patrick
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Prolly get my RMA'ed X9SRL-F board
hah j/k
I tested my build this evening
- 1x CSE815TQ-600CB
- 1x X9DRD-EF
- 2x E5-2670v1
- 16x DIMMs Micron 8Gb 1600Mhz
- 1x SanDisk SSD 64Gb SATA
- 1x Chelsio T480-CR (Dual 10Gb/s NIC)
so the boot hang on BA postcode.
So from supermicro, it said it's memory modules.
I had swapped all dimms, tried differents combination to achieve that
- If i put a DIMM on P2DIMM-H1 i will have the BA postcode hangs
- If i put a DIMM on P2DIMM-H2, i will not have the VGA screen at all...
I was able to install 112 Gb sucessfully (14 DIMMs) and i'm installing Windows 10 just for see i can fo further.
So, my questions are:
- Is it a mainboard issue or processor issue ?
- What do you consider to use for test the system (memory, cpu and general stabiliy ) ? benchs ?
I can try with a different processor in CPU2 (i have 2x others E5-2670). But i will do it maybe tomorrow, i'm tired...
It was too beautiful that would work fine... :-/
- 1x CSE815TQ-600CB
- 1x X9DRD-EF
- 2x E5-2670v1
- 16x DIMMs Micron 8Gb 1600Mhz
- 1x SanDisk SSD 64Gb SATA
- 1x Chelsio T480-CR (Dual 10Gb/s NIC)
so the boot hang on BA postcode.
So from supermicro, it said it's memory modules.
I had swapped all dimms, tried differents combination to achieve that
- If i put a DIMM on P2DIMM-H1 i will have the BA postcode hangs
- If i put a DIMM on P2DIMM-H2, i will not have the VGA screen at all...
I was able to install 112 Gb sucessfully (14 DIMMs) and i'm installing Windows 10 just for see i can fo further.
So, my questions are:
- Is it a mainboard issue or processor issue ?
- What do you consider to use for test the system (memory, cpu and general stabiliy ) ? benchs ?
I can try with a different processor in CPU2 (i have 2x others E5-2670). But i will do it maybe tomorrow, i'm tired...
It was too beautiful that would work fine... :-/
Last edited:
Would a CPU cooler like this prevent any throttling problems? Amazon.com: Supermicro 4U Active CPU Heatsink Cooling for X9 UP/DP Systems SNK-P0050AP4: Computers & Accessories
I'm going to be using a standard desktop case so space isn't a concern.
That's the cooler I have and I don't see any throttling under AIDA64. It stays at a constant 3.0GHz per core over a couple hour duration.
The sad thing is that 68$usd is still $100 Canadian. The massive dumping of oil on the markets is killing my ability to buy new crap.
I think my problem is software because I have noctua 120mm coolers on my chips and under load they never get above 55C but when loading mprime on one thread it should boost to past 3Ghz but they don't the highest I've ever seen is 3ghz.
likely specific to my board, will have to turn off any throttling that I can find but here is what it is:
32 threads loaded, torture test, 1-small FFTs (full super duper load)
Code:
Cpu speed from cpuinfo 2593.00Mhz
True Frequency (without accounting Turbo) 2593 MHz
Socket [0] - [physical cores=8, logical cores=16, max online cores ever=8]
CPU Multiplier 26x || Bus clock frequency (BCLK) 99.73 MHz
TURBO ENABLED on 8 Cores, Hyper Threading ON
Max Frequency without considering Turbo 2692.73 MHz (99.73 x [27])
Max TURBO Multiplier (if Enabled) with 1/2/3/4/5/6 cores is 33x/33x/32x/32x/31x/31x
Real Current Frequency 2723.58 MHz (Max of below)
Core [core-id] :Actual Freq (Mult.) C0% Halt(C1)% C3 % C6 % C7 % Temp VCore
Core 1 [0]: 2723.57 (27.31x) 99.6 0 0 0 0 47 1.1359
Core 2 [1]: 2723.57 (27.31x) 99.6 0 0 0 0 50 1.1359
Core 3 [2]: 2723.57 (27.31x) 99.6 0 0 0 0 50 1.1359
Core 4 [3]: 2723.57 (27.31x) 99.6 0 0 0 0 47 1.1359
Core 5 [4]: 2723.57 (27.31x) 99.6 0 0 0 0 44 1.1359
Core 6 [5]: 2723.57 (27.31x) 99.6 0 0 0 0 45 1.1359
Core 7 [6]: 2723.57 (27.31x) 99.6 0 0 0 0 51 1.1359
Core 8 [7]: 2723.58 (27.31x) 99.6 0 0 0 0 48 1.1359
CPU Multiplier 26x || Bus clock frequency (BCLK) 99.73 MHz
TURBO ENABLED on 8 Cores, Hyper Threading ON
Max Frequency without considering Turbo 2692.73 MHz (99.73 x [27])
Max TURBO Multiplier (if Enabled) with 1/2/3/4/5/6 cores is 33x/33x/32x/32x/31x/31x
Real Current Frequency 2756.32 MHz (Max of below)
Core [core-id] :Actual Freq (Mult.) C0% Halt(C1)% C3 % C6 % C7 % Temp VCore
Core 1 [8]: 2756.31 (27.64x) 99.7 0 0 0 0 50 1.1108
Core 2 [9]: 2756.31 (27.64x) 99.7 0 0 0 0 54 1.1108
Core 3 [10]: 2756.31 (27.64x) 99.7 0 0 0 0 50 1.1108
Core 4 [11]: 2756.31 (27.64x) 99.7 0 0 0 0 50 1.1108
Core 5 [12]: 2756.32 (27.64x) 99.7 0 0 0 0 46 1.1108
Core 6 [13]: 2756.32 (27.64x) 99.7 0 0 0 0 47 1.1108
Core 7 [14]: 2756.32 (27.64x) 99.7 0 0 0 0 52 1.1108
C0 = Processor running without halting64x) 99.7 0 0 0 0 46 1.1108
C1 = Processor running with halts (States >C0 are power saver)
C3 = Cores running with PLL turned off and core cache turned off
C6 = Everything in C3 + core state saved to last level cache
Above values in table are in percentage over the last 1 sec
[core-id] refers to core-id number in /proc/cpuinfo
'Garbage Values' message printed when garbage values are read
Ctrl+C to exit
Code:
Cpu speed from cpuinfo 2593.00Mhz
True Frequency (without accounting Turbo) 2593 MHz
Socket [0] - [physical cores=8, logical cores=16, max online cores ever=8]
CPU Multiplier 26x || Bus clock frequency (BCLK) 99.73 MHz
TURBO ENABLED on 8 Cores, Hyper Threading ON
Max Frequency without considering Turbo 2692.73 MHz (99.73 x [27])
Max TURBO Multiplier (if Enabled) with 1/2/3/4/5/6 cores is 33x/33x/32x/32x/31x/31x
Real Current Frequency 3085.01 MHz (Max of below)
Core [core-id] :Actual Freq (Mult.) C0% Halt(C1)% C3 % C6 % C7 % Temp VCore
Core 1 [0]: 2991.88 (30.00x) 100 0 0 0 0 44 1.1659
Core 2 [1]: 2992.86 (30.01x) 1 2.26 0 0 97 43 1.1659
Core 3 [2]: 2991.96 (30.00x) 100 0 0 0 0 47 1.1659
Core 4 [3]: 2824.34 (28.32x) 0 100 0 0 0 38 1.1659
Core 5 [4]: 2991.93 (30.00x) 100 0 0 0 0 43 1.1659
Core 6 [5]: 2826.80 (28.34x) 0 100 0 0 0 39 1.1659
Core 7 [6]: 2991.93 (30.00x) 100 0 0 0 0 47 1.1659
Socket [1] - [physical cores=8, logical cores=16, max online cores ever=8]
CPU Multiplier 26x || Bus clock frequency (BCLK) 99.73 MHz
TURBO ENABLED on 8 Cores, Hyper Threading ON
Max Frequency without considering Turbo 2692.73 MHz (99.73 x [27])
Max TURBO Multiplier (if Enabled) with 1/2/3/4/5/6 cores is 33x/33x/32x/32x/31x/31x
Real Current Frequency 3056.86 MHz (Max of below)
Core [core-id] :Actual Freq (Mult.) C0% Halt(C1)% C3 % C6 % C7 % Temp VCore
Core 1 [8]: 3056.85 (30.65x) 100 0 0 0 0 50 1.1409
Core 2 [9]: 3010.06 (30.18x) 1 100 0 0 0 46 1.1409
Core 3 [10]: 3056.86 (30.65x) 100 0 0 0 0 47 1.1409
Core 4 [11]: 2888.74 (28.97x) 1 34.9 0 0 65.1 41 1.1409
Core 5 [12]: 3056.85 (30.65x) 100 0 0 0 0 44 1.1309
Core 6 [13]: 2840.23 (28.48x) 1 33.3 0 0 66.7 40 1.1409
Core 7 [14]: 3056.85 (30.65x) 100 0 0 0 0 49 1.1409
Core 8 [15]: 2865.97 (28.74x) 0 100 0 0 0 40 1.1409
C1 = Processor running with halts (States >C0 are power saver)
C3 = Cores running with PLL turned off and core cache turned off
C6 = Everything in C3 + core state saved to last level cache
Above values in table are in percentage over the last 1 sec
[core-id] refers to core-id number in /proc/cpuinfo
'Garbage Values' message printed when garbage values are read
Ctrl+C to exit
Last edited:
Regarding throttling, I was able to dump the RAPL TDP power MSRs using the `power_gov' linux binary from this link:
Intel® Power Governor | Intel® Developer Zone
Using it, it seems that my issue between my evga X79 motherboard and the SuperMicro seems to come down to the short term power duration setting from the BIOS. On the evga it's set to 1s, but on the SuperMicro I get 8.8s! The short and long term TDP settings on the evga is it showing that the BIOS is setting the MSRs based on what I put in, but it seems that the chip just doesn't care (long term TDP is 115W, short term is 138W (which is 1.2x the long term, just as Intel recommends!)) It would seem that the evga setting is so short I never actually see it boost. =(
Intel® Power Governor | Intel® Developer Zone
Using it, it seems that my issue between my evga X79 motherboard and the SuperMicro seems to come down to the short term power duration setting from the BIOS. On the evga it's set to 1s, but on the SuperMicro I get 8.8s! The short and long term TDP settings on the evga is it showing that the BIOS is setting the MSRs based on what I put in, but it seems that the chip just doesn't care (long term TDP is 115W, short term is 138W (which is 1.2x the long term, just as Intel recommends!)) It would seem that the evga setting is so short I never actually see it boost. =(
I've ran my 2x E5-2670 (c2 stepping) at 100% load for 30min+ and not been throttled with 2U Dynatron Active HSF. As I said before I think they were 2.999 / 3.0 / 3.1ghz 100% all the time.
Obviously these will go up to like 8,000 RPM if need be (they didn't). This was with Supermicro board
Obviously these will go up to like 8,000 RPM if need be (they didn't). This was with Supermicro board
I swapped the CPU2 and the result is the same. I did'nt see bent pins (but it's hard to say it precisely).
So i think it's a motherboard issue. I contacted SuperMicro and i will see.
I doubt i will do an RMA to the ebay seller for 16Gb RAM unavailable on a total of 128Gb.. i will see.
You're probably limited to bclk on these. I thought i7z would let me force the multiplier but my mobo won't let me overstep the TDP limit.
They don't overclock at all, locked multi even the bclk straps are locked. BCLK overclocking isn't worth it.
The ones that overclock are some SKU's from 16XX series...like the 1650,1660 and the 1680v2.
Actually I've seen a 4.8GHZ E5-1680 V2 on a overclocking forum and needless to say it was a beast.
What size standoffs did you get?I've got a new case and bunch of plastic stand offs incoming.
Decided to do a test run today before the case arrived. All seems to be fine