- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Does any Intel Gladden Ivy Bridge supports RAM to RAM DMA?
How I can determine support of that technology by my CPU from Linux OS?
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I have the exact the same question. This is a very poorly covered in available documentation.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I think found something that can help you determine if it does or not support : http://stackoverflow.com/questions/37766230/how-to-transfer-data-via-dma-from-ram-to-ram linux - How to transfer data via DMA from RAM to RAM? - Stack Overflow
From what I know all Intel CPUs should support it as is a very common operation.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello, All:
Ronin thank you for your input.
Ficus, could you please provide me with your processor model?
Regards,
Amy.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello, /thread/105465 Ficus:
I just wanted to check on your thread; did the previous input solve your inquiry? If you still require assistance please let me know.
Regards,
Amy.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
No problem didn't solved. It's CPUINFO output
processor : 0vendor_id : GenuineIntelcpu family : 6model : 86model name : Intel(R) Xeon(R) CPU D-1527 @ 2.20GHzstepping : 3microcode : 0x700000ccpu MHz : 2194.870cache size : 6144 KBphysical id : 0siblings : 8core id : 0cpu cores : 4apicid : 0initial apicid : 0
fpu : yesfpu_exception : yes
cpuid level : 20wp : yesflags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdt scp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc ap erfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch ida arat epb xsaveopt pln pts d therm tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 sm ep bmi2 erms invpcid rtm rdseed adx smapbogomips : 4389.74clflush size: 64cache_alignment : 64
address sizes : 46 bits physical, 48 bits virtual
power management:
processor : 1vendor_id : GenuineIntelcpu family : 6model : 86model name : Intel(R) Xeon(R) CPU D-1527 @ 2.20GHzstepping : 3microcode : 0x700000ccpu MHz : 2194.870cache size : 6144 KBphysical id : 0siblings : 8core id : 1cpu cores : 4apicid : 2initial apicid : 2
fpu : yesfpu_exception : yes
cpuid level : 20wp : yesflags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdt scp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc ap erfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch ida arat epb xsaveopt pln pts d therm tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 sm ep bmi2 erms invpcid rtm rdseed adx smapbogomips : 4389.74clflush size: 64cache_alignment : 64
address sizes : 46 bits physical, 48 bits virtual
power management:
processor : 2vendor_id : GenuineIntelcpu family : 6model : 86model name : Intel(R) Xeon(R) CPU D-1527 @ 2.20GHzstepping : 3microcode : 0x700000ccpu MHz : 2194.870cache size : 6144 KBphysical id : 0siblings : 8core id : 2cpu cores : 4apicid : 4initial apicid : 4
fpu : yesfpu_exception : yes
cpuid level : 20wp : yesflags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdt scp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc ap erfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch ida arat epb xsaveopt pln pts d therm tpr_shadow vnmi flexpriority ept ...- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Here is a test how to test if it supports ram to ram:
http://lxr.free-electrons.com/source/drivers/dma/dmatest.c Linux/drivers/dma/dmatest.c - Linux Cross Reference - Free Electrons
and how to test is here: https://www.kernel.org/doc/Documentation/dmaengine/dmatest.txt https://www.kernel.org/doc/Documentation/dmaengine/dmatest.txt
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Ronin,
That test didn't work, because there aren't any visible DMA channels (folder /sys/class/dma/ is empty). I don't know how to determine, it's hardware doesn't support that DMA type or driver doesn't work properly (ioatdma driver was insmoded)
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Ok, I checked why is not working and I think that they changed the kernel and each device has it's own dma channel, assigned dynamically. I think only if you have an old device that uses old dma style will show a dma channel in that directory.
They changed the way memory is assigned, because before, each device driver was build with some default memory allocations (assigned by the driver developer), and because of that 2 devices could end up using the same memory space and BSOD . Now the kernel takes care of that.
There are some benchmarks that show you the transfer speed memory to memory, and the speed is the same as you were copying something from memory or on the memory and I think that operation is done internally, using only the memory controller.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Ronin again thank you for your input, and let's hope this can help /thread/105465 Ficus.
Regards,
Amy.
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page