Are we only using 50% of the resources?

[held]

Are we only using 50% of the resources? When I look at the DE-10 nano specs in Linux I'm getting a whole different picture. Did I overlook something ?

DE-10 Specifications

Code: Select all

>>    800MHz Dual-core ARM >>>Cortex-A9<<< processor
>>    1GB DDR3 SDRAM (32-bit data bus)
    1 Gigabit Ethernet PHY with RJ45 connector
    USB OTG Port, USB Micro-AB connector
    Micro SD card socket
    Accelerometer (I2C interface + interrupt)
    UART to USB, USB Mini-B connector
    Warm reset button and cold reset button
    One user button and one user LED
    LTC 2x7 expansion header

cat /proc/cpuinfo

Code: Select all

processor	: 0
>> model name	: ARMv7 Processor rev 0 (v7l)
BogoMIPS	: 200.00
Features	: half thumb fastmult vfp edsp thumbee neon vfpv3 tls vfpd32 
CPU implementer	: 0x41
CPU architecture: 7
CPU variant	: 0x3
CPU part	: 0xc09
CPU revision	: 0

processor	: 1
model name	: ARMv7 Processor rev 0 (v7l)
BogoMIPS	: 200.00
Features	: half thumb fastmult vfp edsp thumbee neon vfpv3 tls vfpd32 
CPU implementer	: 0x41
CPU architecture: 7
CPU variant	: 0x3
CPU part	: 0xc09
CPU revision	: 0

Hardware	: Altera SOCFPGA
Revision	: 0000
Serial		: 0000000000000000

cat /proc/meminfo

Code: Select all

>> MemTotal:         504232 kB
MemFree:          381288 kB
MemAvailable:     452956 kB
Buffers:            3916 kB
Cached:            72560 kB
SwapCached:            0 kB
Active:            16476 kB
Inactive:          83652 kB
Active(anon):        172 kB
Inactive(anon):    23780 kB
Active(file):      16304 kB
Inactive(file):    59872 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:             0 kB
SwapFree:              0 kB
Dirty:                 4 kB
Writeback:             0 kB
AnonPages:         23652 kB
Mapped:            14728 kB
Shmem:               300 kB
KReclaimable:       6572 kB
Slab:              14740 kB
SReclaimable:       6572 kB
SUnreclaim:         8168 kB
KernelStack:         624 kB
PageTables:          500 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:      252116 kB
Committed_AS:      36236 kB
VmallocTotal:     516096 kB
VmallocUsed:         448 kB
VmallocChunk:          0 kB
Percpu:              144 kB
CmaTotal:              0 kB
CmaFree:               0 kB

[Sterbi]

DE10-Nano have dual core processor, and output of /proc/cpuinfo seems correct to me - you can see two logical processors, processor #0 and processor #1

AFAIK, 1GB RAM is divided into two segments: 512 MB is available for HPS (available for Linux side of Mister) and another 512 MB is dedicated for FPGA so you won't be able to see it from OS side.

[Thedad]

It is interesting, if you run a core and then run "top" to monitor resource usage, it looks like half of the resources is being used.

I think the Cyclone V is 50% an FPGA and 50% an ARM architecture, so that might be it. i.e. when you run a core, you've 50% of the total noted resource for it.

[sardine]

not knowing linux I'm guessing this is why some devs are attempting Hybrid emulation placing cpu emulation on the linux side. I'm just worried that it's design is to handoff to the fpga and linux to idle, using both might increase unexpected heat and load on the combined chip ??

[Sterbi]

When you execute "top" you won't be able to see utilization of FPGA. FPGA itself is not a CPU that can be monitored in such fashion. Therefore running "top" on ARM/Linux side of Mister FPGA would monitor ARM processor and memory dedicated to this ARM.

[FPGA64]

One Arm core is dedicated to running the code to support the cores on the FPGA.

[bbond007]

It is interesting, if you run a core and then run "top" to monitor resource usage, it looks like half of the resources is being used.

Try htop vs top

That utility is a little more clear to understand as far as CPU usage.

Generally speaking, the MiSTer menu consumes 100% of CPU 0, however if you enter the f9 framebuffer console, it scales it itself back to about 10% usage. So the result will be different as far as CPU usage depending on if you test locally via F9 console vs SSH.

not knowing linux I'm guessing this is why some devs are attempting Hybrid emulation placing cpu emulation on the linux side. I'm just worried that it's design is to handoff to the fpga and linux to idle, using both might increase unexpected heat and load on the combined chip ??

You can see that there is VERY little contextual variation in power draw at all, suggesting that CPU temps should also have little variation (source: Grabulosaure) --> https://www.atari-forum.com/viewtopic.p ... 10#p392210

Running the ao486 and FluidSynth or MUNT on the HPS has never caused any (additional) instability issue for me. I live in Florida which gets hot. I don't even use a fan, just a heatsink in my Amiga 1200 case MiSTer.

[held]

DE10-Nano have dual core processor, and output of /proc/cpuinfo seems correct to me - you can see two logical processors, processor #0 and processor #1

AFAIK, 1GB RAM is divided into two segments: 512 MB is available for HPS (available for Linux side of Mister) and another 512 MB is dedicated for FPGA so you won't be able to see it from OS side.

Ahh..that makes much more sense