AWS Lambda CPU Cores

lambda_cpu_cores.md

AWS Lambda

Whats the maximum number of virtual processor cores available in aws lambda

Memory Settings:

Memory: 3008 MB

processor	: 0
vendor_id	: GenuineIntel
cpu family	: 6
model		: 62
model name	: Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
stepping	: 4
microcode	: 0x42a
cpu MHz		: 2799.950
cache size	: 25600 KB
physical id	: 0
siblings	: 2
core id		: 0
cpu cores	: 1
apicid		: 0
initial apicid	: 0
fpu		: yes
fpu_exception	: yes
cpuid level	: 13
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx rdtscp lm constant_tsc rep_good nopl xtopology eagerfpu pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm fsgsbase smep erms xsaveopt
bugs		:
bogomips	: 5600.17
clflush size	: 64
cache_alignment	: 64
address sizes	: 46 bits physical, 48 bits virtual
power management:

processor	: 1
vendor_id	: GenuineIntel
cpu family	: 6
model		: 62
model name	: Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
stepping	: 4
microcode	: 0x42a
cpu MHz		: 2799.950
cache size	: 25600 KB
physical id	: 0
siblings	: 2
core id		: 0
cpu cores	: 1
apicid		: 1
initial apicid	: 1
fpu		: yes
fpu_exception	: yes
cpuid level	: 13
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx rdtscp lm constant_tsc rep_good nopl xtopology eagerfpu pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm fsgsbase smep erms xsaveopt
bugs		:
bogomips	: 5600.17
clflush size	: 64
cache_alignment	: 64
address sizes	: 46 bits physical, 48 bits virtual
power management:

Total CPU

CPU count: 2

def lambda_handler(event, context):
    import multiprocessing
    from subprocess import Popen, PIPE, STDOUT
    print "CPU count: {0}".format(multiprocessing.cpu_count())
    print "=" * 50
    p = Popen("cat /proc/cpuinfo", shell=True, stdin=PIPE, stdout=PIPE, stderr=STDOUT, close_fds=True)
    output = p.stdout.read()
    print output

That's it?
They give you 2 CPUs for 3GB RAM

@gnsx: Yes as from the docs:

At 1,792 MB, a function has the equivalent of 1 full vCPU

So for 3GB you get 2 CPUs. That makes me wonder on what type of machine they are running lambdas behind the scenes. t3.micro?

@gnsx That's correct, as things currently stand you get 2 CPUs for 3GB RAM

@lcb Under the hood they seem to be using a C3 class instances

@saidsef Oh, you are the first one telling me this. Highly interesting. Any source for that info?

I was told by the AWS CS to think of it as EC2 scaling.

As per my experience with lambda now, the more RAM you provision the more IO/Network you get.

I didn't find any compute difference from 1024 onwards though, network has a significant improvement from 1024MB to 3GB.

The better practice is to run multiple lambdas in parallel if there is more parallization required which is a design issue of it's own if lambda is chosen to do this type of work.

@saidsef Oh, you are the first one telling me this. Highly interesting. Any source for that info?

@lcb As per test result, the CPU type - this is based on 3GB RAM - is C3 class instances (they've now been superseded by C5's, and AWS have updated docs accordingly) see here: https://docs.aws.amazon.com/opsworks/latest/userguide/agent-instance.html

Of course, this all subject to memory size, as @gnsx stated - the more RAM you provision the more IO/Network you get and the different instance type your function will deployed on.

Thanks for all the info. In summary Lambdas

• run on C-class type of machines,
• OS is based on Amazon Linux (2) (https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtimes.html),
• You get CPU, Network, I/O power depending on allocated MEM
  • Multithreading (2CPU) you get for >1.8GB RAM

Not that I actually should care about that, but it sheds some light on the Lambda container black box :)

thank you, really helpful!

Seems like now vCPU cores are proportional to the amount of memory configured:

Since Lambda allocates CPU power proportional to the amount of memory provisioned, customers now have access to up to 6 vCPUs. This helps compute intensive applications like machine learning, modelling, genomics, and high-performance computing (HPC) application perform faster.

(https://aws.amazon.com/de/about-aws/whats-new/2020/12/aws-lambda-supports-10gb-memory-6-vcpu-cores-lambda-functions/)