Use of AWS c4.8x large instance mandatory?

[sclimpro]

Hi, it is stated in CW4 that for one of the tasks:

Task: Generate a plot of P (i.e. which script was used) versus sustained bandwidth (i.e. bytes/sec) for these scripts on a c4.8xlarge AWS instance. Note that you may want to leave this till all of your development work is done, so that you don't keep starting and stopping instances. Save the plot as results/pipeline_p_vs_bandwidth.pdf.

So far, I was unable to obtain the $100 from AWS but can access the t2.2xlarge from AWS educate, which is an 8-vCPU, 32GiB machine. Can I use this instance instead? Thanks.

[sclimpro]

By using the t2.2xlarge, I can still observe the trend of increased sustained speed from P1 to P8, and then speedup reduction for P16 and P32.

[filangelos]

According to AWS Educate Terms of Use, we are not entitled to use c4.8xlarge AWS instance and we are restricted to t2 instance types, m4.large, and m4.xlarge.

[m8pple]

My understanding is that "AWS Educate Starter Account" is a different kind of
account, and not approach I was recommending - it's a way of getting an
extra $75, but is not that same as the main way of getting an AWS account
with full access to the instances.

So my recommendation
is still to go here,
register for a normal AWS Account, and apply for the standard $100.

This may well take some time,
but this is why the idea of using AWS was introduced in CW2, well before it
is needed.

You may well need to search your inbox
for the email containing the credit code.

If people only have access to AWS Educate Starter Accounts, then I'm ok with
people producing graphs on the m4.xlarge for this coursework. However, they
will be much less interesting results.

Note that the CW5 and CW6 targets are not available on the list of AWS Educate
Starter Account instances (the ones that @Filangel pointed out above). You can
still do the coursework without starting those instance types, as you
can check the code for correctness, but you'll be flying blind on how fast your
code works on the target machines.

As a meta-note - presumably everyone experiencing this problem was either
not able to, or did not try to, running CW3 on a GPU? I had this as part of
the aim for CW3:

Explore remote GPU instances via AWS

though in order to reduce the formal assessed load of CW3 (as it takes
some people a while) said this:

At this point I'm going to stop requiring you to create graphs,
and trust you to use the techniques you have to understand the
various scaling properties.

With these kinds of exercises where the exact assessment criteria
(criterion?) are known, it's difficult to know what to require versus
just suggesting, as every extra required/mandated assessed step
means the whole thing takes longer when they are not intended
to take that long (if possible).

[davidpasztor]

Is there any way to receive the $100 credit for a normal AWS Account after accidentally registering an AWS Educate Starter Account? By following the recommended steps from CW2, it seems the account registered is an AWS Educate Starter Account and hence only $75 credit was received, which cannot be used for the necessary instances.
I would be grateful for any help on resolving this issue, since it seems a normal AWS Account (which I already have along with the Educate one, but I haven't received the $100 credit) is necessary for the upcoming CWs.