NUMA support on LDOM?

Hi,
I would like to know whether Oracle VM Server for SPARC version 2.0 support NUMA or not?
We're having some problem with JAVA garbage collection (+XX:UseNUMA is enabled) on LDOM running on Sparc T3.
Restarting the LDOM can help resolve the problem temporary.
Thanks.

It's not that Adobe does not support iOS, it's that Apple does not allow third party plugins on the mobile version of Safari.

Similar Messages

AMD cc:NUMA support...

Hi guy's,
If you have a board that supports ccNUMA, turn it on!!!
Boards that support this are:
K8D Master (series) with 2 CPU's
K8D Master3 (series) with 2 CPU's
You need to have min 2 sticks of memory installed, but better is 4, I talk further about 4 sticks.
2 sticks connected to each CPU.
Now you need to go into the BIOS and turn Node (Memory) Interleave OFF!
Do not confuse this with Memory Interleave, that's per CPU and should stay on.
You need to turn this off, because it should not automaticly use any memory that it can find.
Further you need WinXP SP2! (or Win2000 Datacenter or simular)
Check your memory speed, normaly Sandra or something should show a normal speed for your memory.
E.g.:
Normal Dual channel -> PC-2700x2 is about 4800MB/s (just a sample!)
Even with 2 CPU's and 4 sticks installed (2 for echt CPU)
Sandra shows NUMA not supported.
Now edit your boot.ini and add the /PAE switch!
This is very important, as that will turn your Windows to use ccNUMA.
Then it means that every single program will be using it's closest memory.
And share if needed the speed from the other CPU's memory.
Good written programs, multi-threading can use this to improve a lot.
Now check with Sandra memory bench again!
It should go upto 9000MB/s or such.
Also Sandra should show NUMA as being supported.
This should speedup your system under Windows a lot.
Linux can also use this, but I don't know how to check it.
Anyway, it will give your system a whole lot more memory bandwidth to play with.
This only works on the boards listed, as they have memory connected to each CPU.
The K8T-Master or K8N-Master series with 2 CPU's can't do it because they have only memory connected to CPU1.
Windows Vista should have this by default, but I'm not sure, you need to Google on that.

Well a little update, the K8N Master can do NUMA too according to Sandra.
The /PAE switch should be enough.
For other boards that have memory per CPU.
Disable Memory node interleave.
Turn SRAT on in the BIOS.
Edit the BOOT.INI and add the /PAE switch.
Now reboot and NUMA should be working, but you need at least WinXP SP2 or higher.
Beware, if you don't disable Memory Node Interleave, NUMA will not work!!
NUMA does make a big difference, so use it if you can.
The following AMD boards have NUMA full blown:
K8D Master-133 series
K8N Speedster series
And other DUAL AMD CPU systems that have memory connected to each CPU.
The K8N Master seems to support it, but I'm not quite sure what it does with it, maybe somebody can test and find out.

Kernel with numa support

How easy would it be to get this enabled in the arch64 kernel. I suppose i'm talking to the 64bit kernel package manager.
A lot of dual processor AMD system have numa capabilities. And i suspect there will be qute a number of dual processor owners here.
Also, will there really be a performance increase by enabling this?

Hm. My kernel plays numa quite well already.
>_>
<_<
Yeah, I know, I'm no help.

NUMA is disabled in Oracle Linux

Hello
referring to the latest Oracle Linux 6
https://blogs.oracle.com/linux/entry/oracle_rdbms_server_11gr2_pre
This package (as also the previous oracle-validated rpm) sets the numa=off in kernel boot (even when using Oracle UEK, not just Red Hat kernel)
Assuming the latest Oracle RDBMS, i.e. 11g R2, 11.2.0.3 on X86_64 bit-bit Oracle Linux 5.x and 6.x
Can anyone point me to the position re Oracle 11g R2 support of NUMA on Intel/AMD based servers ?
Many thanks.

yurib wrote:
Hello
referring to the latest Oracle Linux 6
https://blogs.oracle.com/linux/entry/oracle_rdbms_server_11gr2_pre
This package (as also the previous oracle-validated rpm) sets the numa=off in kernel boot (even when using Oracle UEK, not just Red Hat kernel)
Assuming the latest Oracle RDBMS, i.e. 11g R2, 11.2.0.3 on X86_64 bit-bit Oracle Linux 5.x and 6.x
Can anyone point me to the position re Oracle 11g R2 support of NUMA on Intel/AMD based servers ?
Many thanks.NUMA is by default is disable from 11gR2 database versions. I.e enableNUMA_support=FALSE is by default from 11.2 database even if NUMA is configured from hardware/os level. Oracle recommends to do complete test n planning before going in production with NUMA enabled at database and os level.
If you still want to enable NUMA in your database from 11.2, follow below note. For check NUMA support at hardware and OS level you need to check with your vendor.
Enable Oracle NUMA support with Oracle Server Version 11.2.0.1 [ID 864633.1]
Also see
Oracle NUMA Usage Recommendation [ID 759565.1]

Possible to use ZFS export from another server to boot/load LDOM?

Have a T1000 with 2 72GB disks, configured as a RAID mirror in firmware, so there's not a lot of available local disk space.
Would like to know if it's possible to use a ZFS volume exported to the base system, to build an LDOM?
Assumes that we upgrade the T1000 to the 10/08 release that supports ZFS for boot devices.
If it's not likely to succeed using exported ZFS volumes, will it work for an exported disk or slice?
Want to use sorage available from an X4500 to ensure ldoms have enough space to live happily in.

Yes, you can use a ZFS volume as a backend for an LDom.
Chapter 5 of the LDoms Admin Guide has some good detail.
http://docs.sun.com/source/820-4894-10/chapter5.html
"The virtual disk backend can be a physical disk, a physical disk slice, a file, a volume from a volume management framework,
such as the Zettabyte File System (ZFS), Solaris Volume Manager (SVM), Veritas Volume Manager (VxVM), or any disk pseudo
device accessible from the service domain."
You don't explicitly need the ZFS boot support in Solaris 10 10/08 (S10U6) as the LDoms support is independent of that - support
for LDoms devices booting off a ZFS volume was added to the LDoms drivers in Solaris 10 5/08 (S10U5).
So you could do the following.
zfs create -V 10G ldoms/domain/test/zdisk0
ldm add-vdsdev /dev/zvol/dsk/ldoms/domain/test/zdisk0 zdisk0@primary-vds0
ldm add-vdisk zdisk0 zdisk0@primary-vds0 ldg1
An then install Solaris on the guest domain using UFS as the root filesystem
[or ZFS if you install Solaris 10 10/08 (S10U6)]

VMware features in Vpshere 4.1 supported by SAP?

Hi, we're currently redesigning our VMware environment and to that end I'd like to hear whether the following VMware features are supported by SAP for production use:
First of all; the farm is based on Vsphere version 4.1 hosting SAP applications/DBs on w2k3 OS..
the use of EVC AMD Opteron Generation 2 (Enhanced VMotion Compatibility (EVC) processor support)?
the use of Wide-VM NUMA Support when using EVC (if that's a possible scenario!)?
the use of ballooning in terms of being able to utilize the ressources on the VM hosts?
the use of Transparent Page Sharing (TPS) for memory management?
the use of memory Compression?
I've been looking into to the specific notes and guides;
Note 1056052 - Windows VMware ESX 3.x or vSphere configuration guidelines.pdf
Note 1409608 - Virtualization on Windows.pdf
SAP-Best-Practices-White-Paper-2009.pdf
But I dont seem to find an answer to the above questions. Does anybody knows of these issues?
/kurt

Hi Kurt,
I already answered you through the support channel, but it's good to repeat the answers for the community:
the use of EVC AMD Opteron Generation 2 (Enhanced VMotion Compatibility (EVC) processor support)?
You can use EVC. Consider that you have to restart your VM in order to activate EVC.
the use of Wide-VM NUMA Support when using EVC (if that's a possible scenario!)?
Wide VM NUMA is an automatic scheduler enhancement and does not conflict with EVC.
the use of ballooning in terms of being able to utilize the ressources on the VM hosts?
We recommend to install the ballooning driver (which comes with the VMware Tools). But it is required that a VM with an SAP Instance reserve 100 % of its memory ("no-memory-overcommitment-policy"). In that case, no real ballooning takes place, but the driver is recommended to be installed for technical reason. See SAP Note 1056052, paragraph 6.
the use of Transparent Page Sharing (TPS) for memory management?
the use of memory Compression?
These two features are absolutely transparent to a Virtual Machine and
therefore they do not conflict with SAP Systems. It is supported to use
them.
Kind regards,
Matthias

Volume as install disk for Guest Domain and Live Upgrade

Hi Folks,
I am new to LDOMs and have some questions - any pointers, examples would be much appreciated:
(1) With support for volumes to be used as whole disks added in LDOM release 1.0.3, can we export a whole LUN under either VERITAS DMP or mpxio control to guest domain and install Solaris on it ? Any gotchas or special config required to do this ?
(2) Can Solaris Live Upgrade be used with Guest LDOMs ? or is this ability limited to Control Domains ?
Thanks

The answer to your #1 question is YES.
Here's my mpxio enabled device.
non-STMS device name STMS device name
/dev/rdsk/c2t50060E8010029B33d16 /dev/rdsk/c4t4849544143484920373730313036373530303136d0
/dev/rdsk/c3t50060E8010029B37d16 /dev/rdsk/c4t4849544143484920373730313036373530303136d0
create the virtual disk using slice 2
ldm add-vdsdev /dev/dsk/c4t4849544143484920373730313036373530303136d0s2 77bootdisk@primary-vds01
add the virtual disk to the guest domain
ldm add-vdisk apps bootdisk@primary-vds01 ldom1
the virtual disk will be imprted as c0d0 which is the whole lun itself.
bind, start ldom 1 and install OS (i used jumpstart) and it partitioned the boot disk c0d0 as / 15GB, swap the remaining space (10GB)
when you run format, print command on both guest and primary domain on this disk you'll see the same slice/size information
Part Tag Flag Cylinders Size Blocks
0 root wm 543 - 1362 15.01GB (820/0/0) 31488000
1 swap wu 0 - 542 9.94GB (543/0/0) 20851200
2 backup wm 0 - 1362 24.96GB (1363/0/0) 52339200
3 unassigned wm 0 0 (0/0/0) 0
4 unassigned wm 0 0 (0/0/0) 0
5 unassigned wm 0 0 (0/0/0) 0
6 unassigned wm 0 0 (0/0/0) 0
7 unassigned wm 0 0 (0/0/0) 0
I havent used DMP but HDLM (Hitachi Dynamic link manager) seems not supported by ldom as i cannot make it work :(
I have no answer on your next question unfortunately.

Warranty service for K9ND Speedster

Hello Good People of the World!
I seem to have a problem and need some input if anyone can offer any.
I sold a K9ND Speedster to a man in Canada. Within a week of its arrival, he writes and says its defective. He contacted MSI Canada and they go over the issue and say it needs an RMA, to contact the seller (Me) and get an RMA and send it back. I contacted MSI and ordered another board for the customer (I aim to please). When the defective item arrived, I shipped the replacement. I shipped the replacement to MSI in California. I was contacted and was told the unit was not defective. They will be shipping it back to me. 5 days ago, the guy I sent a replacement to contacts me and said the second board is defective. I asked him what components he was using to build his system. Everything he told me should work except the power supply. It was a common 300 watt. It had a 20 pin connector. I suggested he contact a qualified tech to help him with his situation because he did not have adequate power. At this point, I figure he has no idea what he's doing. The contacted me 2 days ago and says he brought it to a "Qualified" tech. Had it tested with the now correct power supply and finds it defective.
Can anyone offer some in-site to this problem? I had already replaced a board that was not defective.
Thanks

It Works!!!!
Thanks All
It seems very stable so far.
Here are the Benchmarks:
Processor Arithmatic:
SiSoftware Sandra
Benchmark Results
Dhrystone ALU : 31210 MIPS
Whetstone iSSE3 : 27079 MFLOPS
Results Interpretation : Higher index values are better.
Performance Test Status
Run ID : DYNAMO-C14D7E7B on Thursday, August 02, 2007 at 7:41:11 AM
Platform Compliance : Win32 x86
NUMA Support : Yes
Number of NUMA Nodes : 2
SMP Test : Yes
Total Test Threads : 4
Multi-Core Test : Yes
Cores per Processor : 2
SMT Test : No
Dynamic MP/MT Load Balance : No
Processor Affinity : P0C0T0 P1C0T0 P0C1T0 P1C1T0
System Timer : 2.2GHz
Number of Runs : 64000 / 640
Processor
Model : 2x Dual-Core AMD Opteron(tm) Processor 2214
Speed : 2.21GHz
Model Number : 4420 (estimated)
Cores per Processor : 2 Unit(s)
Type : Dual-Core
L2 On-board Cache : 1MB ECC Synchronous, Write-Back, 16-way set, 64 byte line size
Features
SSE Technology : Yes
SSE2 Technology : Yes
SSE3 Technology : Yes
Supplemental SSE3 Technology : No
SSE4 Technology : No
EMMX - Extended MMX Technology : Yes
SSE4A Technology : No
HTT - Hyper-Threading Technology : No
Performance Tips
Notice 5008 : To change benchmarks, click Options.
Notice 5004 : Synthetic benchmark. May not tally with 'real-life' performance.
Notice 5006 : Only compare the results with ones obtained using the same version!
Tip 2 : Double-click tip or press Enter while a tip is selected for more information about the tip.
Memory Bandwidth:
SiSoftware Sandra
Benchmark Results
RAM Bandwidth Int Buff'd iSSE2 : 6103 MB/s
RAM Bandwidth Float Buff'd iSSE2 : 6543 MB/s
Results Interpretation : Higher index values are better.
Int Buff'd iSSE2 (Integer STREAM) Results Breakdown
Assignment : 5974MB/s
Scaling : 6024MB/s
Addition : 6208MB/s
Triad : 6208MB/s
Data Item Size : 16 byte(s)
Buffering Used : Yes
Offset Displacement Used : Yes
Bandwidth Efficiency : 30% (estimated)
Float Buff'd iSSE2 (Float STREAM) Results Breakdown
Assignment : 6382MB/s
Scaling : 6453MB/s
Addition : 6701MB/s
Triad : 6637MB/s
Data Item Size : 16 byte(s)
Buffering Used : Yes
Offset Displacement Used : Yes
Bandwidth Efficiency : 32% (estimated)
Performance Test Status
Run ID : DYNAMO-C14D7E7B on Thursday, August 02, 2007 at 7:51:06 AM
Platform Compliance : Win32 x86
Memory Used by Test : 512MB
NUMA Support : Yes
Number of NUMA Nodes : 2
SMP Test : Yes
Total Test Threads : 4
Multi-Core Test : Yes
SMT Test : No
Dynamic MP/MT Load Balance : No
Processor Affinity : P0C0T0 P1C0T0 P0C1T0 P1C1T0
System Timer : 2.2GHz
Page Size : 4kB
Use Large Memory Pages : No
NUMA Information
Number of NUMA Nodes : 2
Node 0 - Free Memory : 372MB
Node 1 - Free Memory : 141MB
Features
SSE Technology : Yes
SSE2 Technology : Yes
SSE3 Technology : Yes
Supplemental SSE3 Technology : No
SSE4 Technology : No
EMMX - Extended MMX Technology : Yes
SSE4A Technology : No
HTT - Hyper-Threading Technology : No
Chipset 1
Model : Advanced Micro Devices (AMD) Athlon 64 / Opteron HyperTransport Technology Configuration
Front Side Bus Speed : 2x 1005MHz (2010MHz data rate)
In/Out Width : 16-bit / 16-bit
Maximum Bus Bandwidth : 8040MB/s (estimated)
Logical/Chipset 1 Memory Banks
Bank 0 : 2GB ECC Registered DDR2-SDRAM 5.0-5-5-13 (tCL-tRCD-tRP-tRAS) CR1
Channels : 1
Speed : 2x 315MHz (630MHz data rate)
Width : 128-bit
Memory Controller in Processor : Yes
Cores per Memory Controller : 2 Unit(s)
Maximum Memory Bus Bandwidth : 10080MB/s (estimated)
Chipset 2
Model : Advanced Micro Devices (AMD) Athlon 64 / Opteron HyperTransport Technology Configuration
Front Side Bus Speed : 2x 1005MHz (2010MHz data rate)
In/Out Width : 16-bit / 16-bit
Maximum Bus Bandwidth : 8040MB/s (estimated)
Logical/Chipset 2 Memory Banks
Bank 0 : 2GB ECC Registered DDR2-SDRAM 5.0-5-5-13 (tCL-tRCD-tRP-tRAS) CR1
Channels : 1
Speed : 2x 315MHz (630MHz data rate)
Width : 128-bit
Memory Controller in Processor : Yes
Cores per Memory Controller : 2 Unit(s)
Maximum Memory Bus Bandwidth : 10080MB/s (estimated)
Performance Tips
Notice 5008 : To change benchmarks, click Options.
Notice 5004 : Synthetic benchmark. May not tally with 'real-life' performance.
Notice 5006 : Only compare the results with ones obtained using the same version!
Warning 5400 : Low bandwidth efficiency (advanced). Check memory timings and settings.
Warning 5011 : NUMA nodes memory arrangement inefficient for maximum bandwidth.
Tip 2 : Double-click tip or press Enter while a tip is selected for more information about the tip.

Orion under windows server 2003

Hi i posted a thread about running orion under windows server 2008, i got a response saying that it might not be compatible because orion for windows use an older version.
So now im trying orion under windows server 2003 and i still cant make it run.
orion just finish with no error but also without doing any test, and if i go to the event viewer i can see this error message:
Faulting application orion.exe, version 0.0.0.0, faulting module orion.exe, version 0.0.0.0, fault address 0x0004db7f.
I found another post asking the same question saying that this might be a problem with AMD processors, but it was posted in 2008 so i thought it would be a good idea to repost it and see if anyone know about some solution to this problem.
Thanks

Switched off NUMA support to get a similar problem to go away. Now up against:
SKGFR Returned Error -- Async. write failed on FILE: \\.\X:
OER 1: please look up error in Oracle documentation
rwbase_issue_req: lun_aiorq failed on write
rwbase_run_test: rwbase_issue_req failed
rwbase_run_process: rwbase_run_test failed
rwbase_rwluns: rwbase_run_process failed
orion_thread_main: rw_luns failed
Test error occurred
Orion exiting
Win 2008 Orion 10.2

[875P Neo Series] 875p Neo-FISR (PCB 2.0) Major problems

I just upgraded my system from:
P4 Northwood 3ghz
Gigabyte (obsolute old motherbord)
2 x 512 Hynix pc3200 (400mhz)
ati x800 xt pe 256mb 1,12 ghz
450w NorthQ Silent PSU
Now i was reading a lot of benchmarks and tests and finally i bought the 875Neo-FISR...
So what happened when i did a clean install?
I went from 4500 testscore in 3Dmark05 and down to 3000 - i was expecting to hit over 5000 with the new motherboard.
I have testet the ram (swapped them with some other kingstons and geil same scores roughly)
I got the newest dx and ati drivers etc. i even tried the omega drivers, but no drivers can make up for a difference of 2000 in score result.
Also i read that the bios would now read the prober degrees with the new bios - but my bios allways says 75 degrees celsius and when i boot into windows xp (sp2) its on 60-65 and then goes down to around 37-40 as xp stops loading.
I tried checking for heat problems on the mainboard - and the hottest thing is the cooler on the canterwood chip - 2nd hottest thing is the ati card.
Anyone have any idears? Perhaps its a defect card or any suggestions out there?
Regards.
Forgot to mention that all drivers and utilities etc are ofcause fully updated.
Right now i am running a clean windows xp install with sp2, dx9c and ati drivers + the MSI drivers.

Put my old motherboard back in...
Quote
RAM TEST:
SiSoftware Sandra
Benchmark Results
RAM Bandwidth Int Buff'd iSSE2 : 2912 MB/s
RAM Bandwidth Float Buff'd iSSE2 : 2906 MB/s
Results Interpretation : Higher index values are better.
Int Buff'd iSSE2 (Integer STREAM) Results Breakdown
Assignment : 2915MB/s
Scaling : 2915MB/s
Addition : 2910MB/s
Triad : 2911MB/s
Data Item Size : 16 byte(s)
Buffering Used : Yes
Offset Displacement Used : Yes
Bandwidth Efficiency : 91% (estimated)
Float Buff'd iSSE2 (Float STREAM) Results Breakdown
Assignment : 2910MB/s
Scaling : 2908MB/s
Addition : 2903MB/s
Triad : 2903MB/s
Data Item Size : 16 byte(s)
Buffering Used : Yes
Offset Displacement Used : Yes
Bandwidth Efficiency : 90% (estimated)
Performance Test Status
Run ID : RED on 2. november 2005 at 13:26:36
Memory Used by Test : 512MB
NUMA Support : No
SMP Test : No
Total Test Threads : 1
Multi-Core Test : No
SMT Test : Yes
Dynamic MP/MT Load Balance : No
Processor Affinity : P0C0T0
Page Size : 4kB
Use Large Memory Pages : No
Chipset 1
Model : Silicon Integrated Systems (SiS) SiS648 CPU to PCI Bridge
Front Side Bus Speed : 4x 201MHz (804MHz data rate)
Width : 64-bit
Maximum Bus Bandwidth : 6432MB/s (estimated)
Logical/Chipset 1 Memory Banks
Bank 0 : 512MB DDR-SDRAM 3.0-3-3-8 2CMD
Bank 1 : 512MB DDR-SDRAM 3.0-3-3-8 2CMD
Channels : 1
Speed : 2x 201MHz (402MHz data rate)
Width : 64-bit
Maximum Memory Bus Bandwidth : 3216MB/s (estimated)
Features
(W)MMX Technology : Yes
SSE Technology : Yes
SSE2 Technology : Yes
HTT - Hyper-Threading Technology : Yes
CPU Test
SiSoftware Sandra
Benchmark Results
Dhrystone ALU : 9079 MIPS
Whetstone iSSE2 : 6641 MFLOPS
Results Interpretation : Higher index values are better.
Performance Test Status
Run ID : RED on 2. november 2005 at 13:24:35
NUMA Support : No
SMP Test : No
Total Test Threads : 2
Multi-Core Test : No
SMT Test : Yes
Threads per Core : 2
Dynamic MP/MT Load Balance : No
Processor Affinity : P0C0T0 P0C0T1
Number of Runs : 64000 / 640
Processor
Model : 1x Intel(R) Pentium(R) 4 CPU 3.00GHz
Speed : 3.02GHz
Performance Rating : PR4010 (estimated)
Cores per Processor : 1 Unit(s)
L2 On-board Cache : 512kB ECC Synchronous, ATC, 8-way set, 64 byte line size, 2 lines per sector
Chipset 1
Model : Silicon Integrated Systems (SiS) SiS648 CPU to PCI Bridge
Front Side Bus Speed : 4x 201MHz (804MHz data rate)
Features
SSE2 Technology : Yes
SSE3 Technology : No
SSE4 Technology : No
HTT - Hyper-Threading Technology : Yes
As you can see the ram now perform considerably worse but the cpu is actually beeing used now.
So i reckon i got a defect motherboard.
Sadly the company where i bought it wont switch it with a new one before its been back to the factory and i aint waiting a month on a new board :(
So think ill just use my 14 day return rights and get a asus instead.

Mushkin PC3200 Special 2-2-2 results and help!!!

Just bought the Mushkin Special PC3200 2-2-2 1gb (512x2)
Bios 2.1
Here's some disappointing result:
@400 @400
Turbo Ultra Turbo
PAT - Yes PAT - YES
2-2-3-6 (SPD) 2-2-2-6 (SPD)
@2.65V @2.75v
SANDRA 4491/4507 4592/4587
AIDA 4456/1775 4668/1966
Hyper X 2x256 (CL2) (my previous memory)
@433 (PC3500)
Ultra Turbo
PAT - Yes
2-3-3-7-4
@2.8v
4877/4889 (SANDRA2004)
4956/1698 (AIDA)
AIDA writing speed is much faster...but other result not so good.
Observation:
FSB exactly @800. Previously always o/c by itself.
Problem:
Can't o/c even at 2-2-3-7. tried 433 and post... but could not going into XP. reset..could not go into BIOS....reset..."Double Fault" come out.
Any advice would be appreciated!!!. I only want to oc to DDR433 or 454.

Aaron,
sorry. it's not the 4 or 8. just rebooted using 8 and PAT still off.
SiSoftware Sandra (result)
Int Buffered iSSE2 (Integer STREAM) Results Breakdown
Assignment : 4830MB/s
Scaling : 4835MB/s
Addition : 4921MB/s
Triad : 5039MB/s
Data Item Size : 16-bytes
Buffering Used : Yes
Offset Displacement Used : Yes
Bandwidth Efficiency : 71% (estimated)
Float Buffered iSSE2 (Float STREAM) Results Breakdown
Assignment : 4841MB/s
Scaling : 4821MB/s
Addition : 5033MB/s
Triad : 5037MB/s
Data Item Size : 16-bytes
Buffering Used : Yes
Offset Displacement Used : Yes
Bandwidth Efficiency : 71% (estimated)
Test Status
Memory Used by Test : 512MB
NUMA Support : No
SMP Test : No
Total Test Threads : 1
SMT Test : Yes
Number of Test SMT (per CPU) : 2
Dynamic MP/MT Load Balance : No
Processor Affinity : No
Chipset 1
Model : Micro-Star International Co Ltd (MSI) 82875P Memory Controller Hub
Front Side Bus Speed : 4x 217MHz (868MHz data rate)
Width : 64-bit
Maximum Bus Bandwidth : 6944MB/s (estimated)
Logical/Chipset 1 Memory Banks
Bank 0 : 512MB DDR-SDRAM 2.0-3-2-5CL 1CMD
Bank 1 : 512MB DDR-SDRAM 2.0-3-2-5CL 1CMD
Channels : 2
Speed : 2x 217MHz (434MHz data rate)
Width : 64-bit
Performance Acceleration Technology : No
Maximum Memory Bus Bandwidth : 6944MB/s (estimated)
Features
MMX Technology : Yes
SSE Technology : Yes
SSE2 Technology : Yes
SSE3 Technology : No
HTT - Hyper-Threading Technology : Yes
i think its the RAM. Once I overclock, it disable PAT.
I found this in the Mushkin website, any idea that it means? (the latter part)
USER TIPS: The NEO-LSR automatically overclocks both the FSB and DDR frequencies anywhere from 2 to 4 mHz. i.e. FSB 200 reads 202, DDR 400 reads 404, etc. An effect of this is that the board sets the CAS latency to 2.5, even though it will display 2. This applies to all modules slower than the PC3200 222 Specials (with less than 2.75V); the rest will read CAS correctly. To compensate, the FSB should be underclocked so as not to encourage unstable memory speeds.

Need stable settings dont care about overclocking

hello if anyone has msi 865pe ls board
with p4 2.8 800 with ht and 1 gig of corsair 3200 ram pls help me get my system stable. I got it on default but keeps booting me to desktop when playing games. I have 80 gig hard drive and geforce3 ti500 and 5 fans and memorex dvd rom.

Have seen several posts where people are having problems with Corsair ram. I have used a lot of Corsair on older boards & overclocked with no problem. I would try raising the voltage one notch & see if that helps. OCing is not an exact science and what works on one board, may not work on another. More info in your post would be very helpful also.
MSI 865PE NEO2 Model MS 6728 Ver.1
Pentium 4 2.8 ghz 1.525 Volt BX80532PG2800DSL6WT Malay (9-09-03)
Asylum Nvidia 5600 FX video 8x
Realtek 8139 fast NIC (DSL Bellsouth)
HT- enabled
2- 120gb Seagate sata harddrives
pat/mat (whatever) set at fast
fsb- set manually @ 203
2- 512mb ddr pc3200 ram in slot 1&3 "Elixer"
SiSoftware Sandra
Bandwidth Efficiency : 74% (estimated)
Memory Used by Test : 512MB
NUMA Support : No
SMP Test : No
Total Test Threads : 1
SMT Test : Yes
Number of Test SMT (per CPU) : 2
Dynamic MP/MT Load Balance : No
Processor Affinity : No
Chipset 1
Model : Micro-Star International Co Ltd (MSI) 82865G/PE/P, 82848P DRAM
Controller / Host-Hub Interface
Front Side Bus Speed : 4x 203MHz >>>"812MHz data rate"Width : 64-bit
Maximum Bus Bandwidth : 6496MB/s (estimated)
Bank 0 : 512MB DDR-SDRAM 3.0-3-3-6CL 1CMD
Bank 1 : 512MB DDR-SDRAM 3.0-3-3-6CL 1CMD
Channels : 2
Speed : 2x 203MHz (406MHz data rate)
Width : 64-bit
Performance Acceleration Technology : Yes
Maximum Memory Bus Bandwidth : 6496MB/s (estimated)
Features
MMX Technology : Yes
SSE Technology : Yes
SSE2 Technology : Yes
SSE3 Technology : No
HTT - Hyper-Threading Technology : Yes
Liteon 52x24x52 CDRW
Hitachie GD5000 DVD
replaced factory Intel chip cooling & CPU cooling with Swiftech Heatsinks
Rock Stable Win2k Pro
Award 1.7 Bios

Athlon64 & Suse9 for AMD64 edition

Hi there...
I encounter Serious Problemstrying to install Suse9 for AMD64.
Environment: MSI K8T Neo, Athlon64 3000+, Adaptec 29160, 1Gig Corsair PC400 Ram, Nvidia Gforce 4600.
I encounter a Kernel Panic on trying to load the kernel similar to this :
Code:Bad RIP value
RIP [000000008010ddbe] RSP [ffffffff804ddfc8]
Kernel panic: Attempted to kill the idle task!
In idle task-no syncing
Unable to handle paging request
A friend told me i need a bios update . well i run on 1.2 thats the newest... any suggestions ?
THX in advance

We had the same problem and figured out that everything works fine when using the kernel paramter numa=off. After several hours of work we managed to get the system running with NUMA support by replacing the RAM modules.
Unfortunately, the user guide doesn't say anything about the relationship between the memory banks and the processors. In fact, it seems like that the 6 banks belong to CPU 1 while the 4 banks belong to CPU 2. When using NUMA, it is important that every CPU has its own memory. Thus, we distributed our 4 memory modules to both CPUs and everything worked fine.
I hope this helped you.
Bye,
Arthur.

P4 2,4GHZ + PT880 Neo

Hey!
I bought the PT880 Neo for my P4 2,4 (The old one without HT!). But my Performance is Bad!
Motherboard Properties:
Motherboard ID 63-1125-009999-00101111-040201-VIA_P6$PT800MSIA7008VMS V1.1 112503
Motherboard Name MSI PT880 Neo-SR (MS-7008)
Front Side Bus Properties:
Bus Type Intel NetBurst
Bus Width 64-bit
Real Clock 133 MHz (QDR)
Effective Clock 533 MHz
Bandwidth 4267 MB/s
Memory Bus Properties:
Bus Type DDR SDRAM
Bus Width 64-bit
Real Clock 200 MHz (DDR)
Effective Clock 400 MHz
Bandwidth 3200 MB/s
Chipset Bus Properties:
Bus Type VIA V-Link
Bus Width 8-bit
Real Clock 67 MHz (ODR)
Effective Clock 533 MHz
Bandwidth 533 MB/s
Motherboard Physical Info:
CPU Sockets/Slots 1
Expansion Slots 5 PCI, 1 AGP
RAM Slots 4 DIMM
Integrated Devices Audio
Form Factor ATX
Motherboard Size 240 mm x 300 mm
Motherboard Chipset PT880
Extra Features CoreCell
Are those results correct? I got the newest drivers and Updates, does the Processor match with the Motherboard?
Please Help me, even Age of Mythology Lags!!

Integer iSSE2: 14918it/s
Float iSSE2: 18662it/s
Run ID: LUKAS on Montag, 1. März 1999 at 18:25:18
NUMA Support: No
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Rendered Image Size: 640x480
Model: Intel(R) Pentium(R) 4 CPU 2.40GHz
Speed: 2.41GHz
Performance Rating: PR2651 (estimated)
Type: Standard
L2 On-board Cache: 512kB ECC Synchronous ATC (8-way sectored, 64
byte line size)
Model: VIA Technologies Inc PT880 CPU to PCI Bridge
Front Side Bus Speed: 4x 134MHz (536MHz data rate)
MMX Technology: Yes
SSE Technology: Yes
SSE2 Technology: Yes
SSE3 Technology: No
HTT - Hyper-Threading Technolo:No
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
>>
>
Combined Index: 78562 operation(s)/min
Endurance Factor: 8.8
512B Files Test: 178674 operation(s)/min
32kB Files Test: 8336 operation(s)/min
256kB Files Test: 3353 operation(s)/min
2MB Files Test: 1160 operation(s)/min
Run ID: LUKAS on Montag, 1. März 1999 at 18:25:38
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Read Performance: 1036698 operation(s)/min (8639 kb/sec, 49x)
Write Performance: 62069 operation(s)/min (517 kb/sec, 2x)
Delete Performance: 541494 operation(s)/min
Combined Index: 178674 operation(s)/min
Read Performance: 7351 operation(s)/min (3921 kb/sec, 22x)
Write Performance: 7867 operation(s)/min (4196 kb/sec, 23x)
Delete Performance: 492451 operation(s)/min
Combined Index: 8336 operation(s)/min
Read Performance: 3119 operation(s)/min (13308 kb/sec, 75x)
Write Performance: 2839 operation(s)/min (12113 kb/sec, 68x)
Delete Performance: 449099 operation(s)/min
Combined Index: 3353 operation(s)/min
Read Performance: 1129 operation(s)/min (38537 kb/sec, 218x)
Write Performance: 907 operation(s)/min (30959 kb/sec, 175x)
Delete Performance: 578653 operation(s)/min
Combined Index: 1160 operation(s)/min
Operating System Disk Cache Us:No
Use Overlapped I/O: No
Test File Size: 32MB
Repeated Sector ReWrite: 2391 kB/s
Sequential Sector Write: 1871 kB/s
Random Sector Write: 168 kB/s
Total Size: 38GB
Free Space: 25GB, 100%
Cluster Size: 4kB
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
Notice 5207: Consider using the File System Benchmark for
non-Flash devices.
Notice 5900: Endurance factor can only be used on the same
type of device (SLC or MLC).
Tip 11: Use the 'Switch Chart Type' button to switch
between Detailed and Combined charts.
Tip 5202: Use cache on to measure Windows performance.
>
Combined Index: 23531 operation(s)/min
Endurance Factor: 7.1
512B Files Test: 8160 operation(s)/min
32kB Files Test: 51754 operation(s)/min
256kB Files Test: 10981 operation(s)/min
2MB Files Test: 1549 operation(s)/min
Run ID: LUKAS on Montag, 1. März 1999 at 18:25:52
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Read Performance: 5810 operation(s)/min (48 kb/sec, 0x)
Write Performance: 20752 operation(s)/min (173 kb/sec, 0x)
Combined Index: 8160 operation(s)/min
Read Performance: 81687 operation(s)/min (43566 kb/sec, 247x)
Write Performance: 33398 operation(s)/min (17812 kb/sec, 101x)
Combined Index: 51754 operation(s)/min
Read Performance: 12906 operation(s)/min (55066 kb/sec, 312x)
Write Performance: 8974 operation(s)/min (38289 kb/sec, 217x)
Combined Index: 10981 operation(s)/min
Read Performance: 1645 operation(s)/min (56149 kb/sec, 319x)
Write Performance: 1423 operation(s)/min (48572 kb/sec, 275x)
Combined Index: 1549 operation(s)/min
Operating System Disk Cache Us:No
Use Overlapped I/O: No
Test File Size: 32MB
Repeated Sector ReWrite: 2914 kB/s
Sequential Sector Write: 2356 kB/s
Random Sector Write: 261 kB/s
Total Size: 20GB
Free Space: 12GB, 100%
Cluster Size: 16kB
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
Notice 5207: Consider using the File System Benchmark for
non-Flash devices.
Notice 5900: Endurance factor can only be used on the same
type of device (SLC or MLC).
Tip 11: Use the 'Switch Chart Type' button to switch
between Detailed and Combined charts.
Tip 5202: Use cache on to measure Windows performance.
>
Combined Index: 20375 operation(s)/min
Endurance Factor: 7.9
512B Files Test: 21172 operation(s)/min
32kB Files Test: 28267 operation(s)/min
256kB Files Test: 9737 operation(s)/min
2MB Files Test: 1387 operation(s)/min
Run ID: LUKAS on Montag, 1. März 1999 at 18:26:01
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Read Performance: 22702 operation(s)/min (189 kb/sec, 1x)
Write Performance: 19229 operation(s)/min (160 kb/sec, 0x)
Combined Index: 21172 operation(s)/min
Read Performance: 25500 operation(s)/min (13600 kb/sec, 77x)
Write Performance: 33763 operation(s)/min (18007 kb/sec, 102x)
Combined Index: 28267 operation(s)/min
Read Performance: 9124 operation(s)/min (38929 kb/sec, 221x)
Write Performance: 10827 operation(s)/min (46195 kb/sec, 262x)
Combined Index: 9737 operation(s)/min
Read Performance: 1416 operation(s)/min (48333 kb/sec, 274x)
Write Performance: 1345 operation(s)/min (45909 kb/sec, 260x)
Combined Index: 1387 operation(s)/min
Operating System Disk Cache Us:No
Use Overlapped I/O: No
Test File Size: 32MB
Repeated Sector ReWrite: 3074 kB/s
Sequential Sector Write: 2291 kB/s
Random Sector Write: 236 kB/s
Total Size: 19GB
Free Space: 14GB, 100%
Cluster Size: 16kB
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
Notice 5207: Consider using the File System Benchmark for
non-Flash devices.
Notice 5900: Endurance factor can only be used on the same
type of device (SLC or MLC).
Tip 11: Use the 'Switch Chart Type' button to switch
between Detailed and Combined charts.
Tip 5202: Use cache on to measure Windows performance.
>>
>
Error: Cannot create/write to drive! Check
permissions.
>
Drive Index: 43MB/s
Run ID: LUKAS on Montag, 1. März 1999 at 18:26:11
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Operating System Disk Cache Us:No
Use Overlapped I/O: Yes
IO Queue Depth: 4 request(s)
Test File Size: 511MB
File Server Optimised: No
Buffered Read: 40 MB/s
Sequential Read: 52 MB/s
Random Read: 34 MB/s
Buffered Write: 52 MB/s
Sequential Write: 49 MB/s
Random Write: 30 MB/s
Average Access Time: 10 ms (estimated)
Drive Type: Hard Disk
Total Size: 38GB
Free Space: 25GB, 100%
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
Notice 5209: Consider using the Removable Storage/Flash
Benchmark for Flash devices.
Tip 5202: Use cache on to measure Windows performance.
>
Drive Index: 49MB/s
Run ID: LUKAS on Montag, 1. März 1999 at 18:27:38
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Operating System Disk Cache Us:No
Use Overlapped I/O: Yes
IO Queue Depth: 4 request(s)
Test File Size: 511MB
File Server Optimised: No
Buffered Read: 40 MB/s
Sequential Read: 56 MB/s
Random Read: 39 MB/s
Buffered Write: 55 MB/s
Sequential Write: 56 MB/s
Random Write: 37 MB/s
Average Access Time: 8 ms (estimated)
Drive Type: Hard Disk
Total Size: 20GB
Free Space: 12GB, 100%
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
Notice 5209: Consider using the Removable Storage/Flash
Benchmark for Flash devices.
Tip 5202: Use cache on to measure Windows performance.
>
Drive Index: 43MB/s
Run ID: LUKAS on Montag, 1. März 1999 at 18:28:54
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Operating System Disk Cache Us:No
Use Overlapped I/O: Yes
IO Queue Depth: 4 request(s)
Test File Size: 511MB
File Server Optimised: No
Buffered Read: 40 MB/s
Sequential Read: 49 MB/s
Random Read: 35 MB/s
Buffered Write: 55 MB/s
Sequential Write: 49 MB/s
Random Write: 34 MB/s
Average Access Time: 8 ms (estimated)
Drive Type: Hard Disk
Total Size: 19GB
Free Space: 14GB, 100%
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
Notice 5209: Consider using the Removable Storage/Flash
Benchmark for Flash devices.
Tip 5202: Use cache on to measure Windows performance.
Benchmark Results >
RAM Bandwidth Int Buffered iSS:2913MB/s
RAM Bandwidth Float Buffered i:2929MB/s
Assignment: 2895MB/s
Scaling: 2893MB/s
Addition: 2935MB/s
Triad: 2932MB/s
Data Item Size: 16-bytes
Buffering Used: Yes
Offset Displacement Used: Yes
Bandwidth Efficiency: 91% (estimated)
Assignment: 2907MB/s
Scaling: 2911MB/s
Addition: 2952MB/s
Triad: 2947MB/s
Data Item Size: 16-bytes
Buffering Used: Yes
Offset Displacement Used: Yes
Bandwidth Efficiency: 91% (estimated)
Run ID: LUKAS on Montag, 1. März 1999 at 18:30:19
Memory Used by Test: 256MB
NUMA Support: No
SMP Test: No
Total Test Threads: 1
SMT Test: No
Dynamic MP/MT Load Balance: No
Processor Affinity: No
Page Size: 4kB
Use Large Memory Pages: No
Model: VIA Technologies Inc PT880 CPU to PCI Bridge
Front Side Bus Speed: 4x 134MHz (536MHz data rate)
Width: 64-bit
Maximum Bus Bandwidth: 4288MB/s (estimated)
Bank 0: 16MB DDR-SDRAM 3.0-3-3-8CL 1CMD
Bank Interleave: 4-way
Speed: 2x 201MHz (402MHz data rate)
Width: 64-bit
Maximum Memory Bus Bandwidth: 3216MB/s (estimated)
MMX Technology: Yes
SSE Technology: Yes
SSE2 Technology: Yes
SSE3 Technology: No
HTT - Hyper-Threading Technolo:No
Notice 5008: To change benchmarks, click Options.
Notice 5004: Synthetic benchmark. May not tally with 'real-
life' performance.
Notice 5006: Only compare the results with ones obtained
using the same version!
>>
Combined Index: 6911MB/s
Speed Factor: 14.7
2kB Blocks: 25777MB/s
4kB Blocks: 26261MB/s
8kB Blocks: 21176MB/s
16kB Blocks: 18821MB/s
32kB Blocks: 18894MB/s
64kB Blocks: 18973MB/s
128kB Blocks: 18988MB/s
256kB Blocks: 17590MB/s
512kB Blocks: 14235MB/s
1MB Blocks: 1788MB/s
4MB Blocks: 1805MB/s
16MB Blocks: 1791MB/s
64MB Blocks: 1789MB/s
256MB Blocks: 1786MB/s
Data Item Size: 16-bytes
Buffering Used: No
Offset Displacement Used: Yes

AMD chipsets?

I'm looking into a few board i was just wondering which one is better AMD-8131 or the AMD-8151? This board has cough my eye what do you all think? Any running a MSI K8T Master2 if so how are they?

I'm also running a K8T Master2, fast and stable here. I like it but I've been looking at a Tyan board with AMD8131, mainly for the NUMA support. But I like this board, it's nice.

NUMA support on LDOM?

Similar Messages

Maybe you are looking for