By: Joe Chang (jchang6.delete@this.Xyahoo.com), November 9, 2008 8:11 pm
Room: Moderated Discussions
Michael S (already5chosen@yahoo.com) on 11/9/08 wrote:
---------------------------
>Joe Chang (jchang6@Xyahoo.com) on 11/9/08 wrote:
>---------------------------
>>Inquirer points out the SPEC Power results posted by AMD for Shanghai and Penryn (Xeon L5420) both at 2.5GHz.
>>AMD results for Shanghai is 731 versus 561 for Xeon.
>>SPEC Power results posted by others for the L5420 were higher, the SuperMicro results
>>being 990, but not as close a system match as the 2 AMD results (Windows instead of Linux, and BEA vs Oracle).
>>Now SPEC Power is not my field, as I have never run it. Inquirer points out the
>>different memory config. However in reading through the reports, AMD has the Xeon
>>L5420 system power at 179, 197 and 279W for idle, 10% and 100% respectively. The
>>SuperMicro result has 97, 108 and 174W respectively. I do not think 2 DIMMs will
>>make up 100W, The difference in the SuperMicro result includes settings:
>>Adjacent Sector Prefetcher disabled in BIOS.
>>C1 Enhanced Mode enabled in BIOS.
>>3pin Server Mode enabled.
>
>http://www.supermicro.com/products/system/2U/6025/SYS-6025W-NTR+.cfm
>http://www.supermicro.com/products/system/1U/6015/SYS-6015C-MT.cfm
>>
>
>Supermicro 6025W-NTR+ tested by AMD and Supermicro 6015C-MTB tested by Supermicro are entirely different machines.
>
>The first one is based on 5400 Seaburg chipset+FB-DIMM memory. 5400 was originally
>developed for high-end workstations. It performs pretty in servers but power consumption
>is not its strong point. Also it performs best with 4-8 DIMMs installed while in
>benchmarked configuration there were just 2 DIMMs.
>
>The second server is based on relatively new 5100MCH chipset a.k.a. San Clemente
>paired with registered DDR2 DIMMs. This chipset is designed for low end and has
>limited* memory capacity. In theory it has lower bandwidth than its bigger brothers
>Blackford, Greencreek and Seaburg. In practice,as long as it doesn't hit a capacity
>wall, San Clemente appears to perform on par with them. And when contested against
>them at 2 DIMMs vs 2 DIMMs level one can expect that it would actually outperform these quad-channel gears.
>
>Of course, there are other differences between two systems, e.g. 6025 has 700W PSU vs 520W PSU on 6015.
>
>-----
>* - San Clemente memory capacity is limited relatively to Intel's high end FB-DIMM
>based chipsets and relatively to high-end 2S Opterons, but relatively to capacity
>of previous Intel low end server chipsets 6-8 dual-rank R-DIMMS supported by San Clemente is a big leap forward.
>
so does the chipset 5400/5000 vs 5100, 4 FB-DIMM vs 2 DDR, and the 700W vs 520W power supply contribute nearly 100W power consumed? or is a significant chunk in the power setting? and yes, Shanghai performance looks decent, can't wait to see SPEC CPU int, TPC-C/E and H numbers
---------------------------
>Joe Chang (jchang6@Xyahoo.com) on 11/9/08 wrote:
>---------------------------
>>Inquirer points out the SPEC Power results posted by AMD for Shanghai and Penryn (Xeon L5420) both at 2.5GHz.
>>AMD results for Shanghai is 731 versus 561 for Xeon.
>>SPEC Power results posted by others for the L5420 were higher, the SuperMicro results
>>being 990, but not as close a system match as the 2 AMD results (Windows instead of Linux, and BEA vs Oracle).
>>Now SPEC Power is not my field, as I have never run it. Inquirer points out the
>>different memory config. However in reading through the reports, AMD has the Xeon
>>L5420 system power at 179, 197 and 279W for idle, 10% and 100% respectively. The
>>SuperMicro result has 97, 108 and 174W respectively. I do not think 2 DIMMs will
>>make up 100W, The difference in the SuperMicro result includes settings:
>>Adjacent Sector Prefetcher disabled in BIOS.
>>C1 Enhanced Mode enabled in BIOS.
>>3pin Server Mode enabled.
>
>http://www.supermicro.com/products/system/2U/6025/SYS-6025W-NTR+.cfm
>http://www.supermicro.com/products/system/1U/6015/SYS-6015C-MT.cfm
>>
>
>Supermicro 6025W-NTR+ tested by AMD and Supermicro 6015C-MTB tested by Supermicro are entirely different machines.
>
>The first one is based on 5400 Seaburg chipset+FB-DIMM memory. 5400 was originally
>developed for high-end workstations. It performs pretty in servers but power consumption
>is not its strong point. Also it performs best with 4-8 DIMMs installed while in
>benchmarked configuration there were just 2 DIMMs.
>
>The second server is based on relatively new 5100MCH chipset a.k.a. San Clemente
>paired with registered DDR2 DIMMs. This chipset is designed for low end and has
>limited* memory capacity. In theory it has lower bandwidth than its bigger brothers
>Blackford, Greencreek and Seaburg. In practice,as long as it doesn't hit a capacity
>wall, San Clemente appears to perform on par with them. And when contested against
>them at 2 DIMMs vs 2 DIMMs level one can expect that it would actually outperform these quad-channel gears.
>
>Of course, there are other differences between two systems, e.g. 6025 has 700W PSU vs 520W PSU on 6015.
>
>-----
>* - San Clemente memory capacity is limited relatively to Intel's high end FB-DIMM
>based chipsets and relatively to high-end 2S Opterons, but relatively to capacity
>of previous Intel low end server chipsets 6-8 dual-rank R-DIMMS supported by San Clemente is a big leap forward.
>
so does the chipset 5400/5000 vs 5100, 4 FB-DIMM vs 2 DDR, and the 700W vs 520W power supply contribute nearly 100W power consumed? or is a significant chunk in the power setting? and yes, Shanghai performance looks decent, can't wait to see SPEC CPU int, TPC-C/E and H numbers