Are you talking about this?
RandomX: added cache QoS support #1774
https://github.com/xmrig/xmrig/pull/1774
XMRig v6.3.0 added new experimental feature (cache QoS)
https://www.reddit.com/r/MoneroMining/comments/hsscux/xmrig_v630_added_new_experimental_feature_cache/

I'm currently testing the new version. So far i could not see any significant change - I'll post here when i update the script for the new version. The change will be automatic, i.e. you won't have to do anything (but like i said, don't expect a higher hashrate).

Yes, in principle it works the same. However, azure is very quick to ban people who are using the free credits from trial-accounts. Also, what will you do after the trial runs out? There's just no long term perspective.
What i mean to say is: Just having fun with a single trial account is probably OK. But if you find yourself planning to sign up for more and more trial-accounts just to abuse the free credits, you are going too far.

First off: please don't use the free credits from trial accounts. You'll need a MSDN-subscription to do any reasonable mining.
My script includes a autotune-feature: if the VM is instantiated on a physical CPU where a lot of neighboring VMs are idle, the script can use the CPU's cache more intensively, resulting in a better hashrate. If you use a VM with many cores, your chances of having neighboring, idle VMs is less (since there are less cores for neighboring VMs available). That's why the F2-type works best (if you get different results, of course i'd love to hear more about it).

The difference between dedicated and low priority-nodes is that the low-priority-nodes are a lot cheaper (low-priority-nodes may be restarted by azure at any time, however, for our mining-application this does not matter).

Since you want to burn your 200$ as fast as possible (i assume you are using a trial account), you should create as many vms as possible. A pool with 10 low priority F2 VMs (20 cores) will consume about 150$ per month, so you'll need more than that. You'll probably have to request a quota increase from the azure-support.

Ok,  but is there any reason to think that verus will surpass e.g. Monero anytime soon in popularity? I'm sorry for asking like this - when i started out i tried to adapt my script to different currencies. Looking back, i feel that the effort was in vain because in fact only mining monero is worthwhile.

Basically since i started doing this some years ago Monero (i.e. randomX) has been the best option for mining with azure. I just skimmed over the homepage of veruscoin and it seems to focus on similar qualities like monero (privacy, cpu-friendliness). However, Monero is matured and much bigger than veruscoin - so until that changes i will recommend mining Monero (and i will mine it myself, too).

I searched a bit on the internet and i found this post:
https://bitcointalk.org/index.php?topic=5210557.0
-> This guy assumes that the FPGA-people will also try to create a FPGA-miner for Verushash 2.1, which will make CPU-mining unprofitable again - do you have any more information about this?

With the 150$ monthly credit from a MSDN-Enterprise-subscription you'll get around 8kH/s mining Monero, which will net you (at the time of writing) around 15$ in crypto-currency per Month.
I like to view this as my retirement fund. I'm doing this long-term, so even a little amount will accumulate to a sizable sum after some time. Also, i expect the exchange-rate for crypto to be _way_ higher by the time i retire...

Thanks for the info. But where did you get the prices from? If i check the pricing i get the following numbers:
F2s VM, low priority batch

Central Korea:
$0.0204/hour

East US
$0.0199/hour

(note that there'll be some cost for the storage on top of that).

And another update. the scripts will now use XMRig version 5.5.3. As always, the script will automatically update itself after some time.

Sorry, i don't have anything for AWS. This is mainly because as far as i know there is no legitimate source of free credits for AWS... May i ask you where you get your AWS-credit from? Or are you using a trial-account?

FYI: another update: the scripts will now use XMRig version 5.5.0. There's no need to do anything, the script will automatically use the new version after some time.

FYI: I have updated the scripts to use version 5.2.1 of XMRig. There's no need to do anything, the script will automatically use the new version after some time.

You'll get an email saying that they made the change (internally this is handled as a support-request). From the replies i got i have to say that azure takes the support-requests quite seriously and they are dedicated to resolve them in a timely manner.

FYI: I requested an increase of the quota for low-priority cores (20)  for one of my batch-accounts. It took some days, but then the quota was increased without any problem.
Requesting the quota-increase is very easy. If you go to your batch-account, there's the item "quotas" on the left pane. After clicking it it will show you your current quota-settings, and there is also a button "Request quota increase". You can just click it and fill the following form to get the increased quota (processing the request will likely take some time, though). For most cases 20 low-priority cores should be enough. I haven't tried it, but i guess if you are asking for a much higher quota azure might actually ask what you are planning to do...

It's probably too early to say, but it seems the profit has increased a bit - according to cryptunit 8kH/s will net you ca. 1$ per day in XMR.

In my account all newly created batch-accounts/Pools now have a quota of 0 - probably that's why you got the message about the CPU core limit.

My existing batch-accounts/pools still have their original quotas - If you have a working batch-account it's probably a good idea to keep re-using that account for now.

I will contact the azure-support and check if they increase the quota on request.

The monero-RandomX-fork has gone live... everything went smoothly as far as i can tell.
Still, it's probably a good idea to check the output of your azure-pools - please write here or pm me if you are having any problems.

Hi everybody,

i have activated the changes. If you experience any problems please contact me (post here or send me a pm).

Hi everybody,

I have almost completed the preparation for the Monero RandomX-fork. The good news: you probably won't have to do anything (if you mine something else than monero your action may be required - see the technical details below).

Technical details:
* I am still running some tests, but i will activate my changes 1-2 days after this post (i will make another post here then).
* I have changed the scripts to use xmrig instead of xmr-stak. xmrig makes a smooth change to randomX much easier and i hardly found any difference in the hashrate.
* Unfortunately, however, xmrig manages the different cryptonight-variants differently than xmr-stak. The script will try to adapt the settings to make things work, though. If you are mining something else than Monero: please check your hashrate and tell me if there is any problem after the change in the scripts (like i said i will post here again in 1-2 days when i make the change).
* In my tests i have found that the F2-type of VMs are the most  cost-effective for running randomX. This means that there will be no need to change the pool-settings for the fork (currently F2 is also the most cost-effective for mining monero). You can expect a hashrate of ca. 400H/s per used core. If you have a MSDN-Enterprise subscription this should net you 7-8KH/s (whether this is a lot or not we'll only know after the fork).

You can have a look here:
https://bitcointalk.org/index.php?topic=3444898.msg47841140#msg47841140
However, as MATHReX already said: if you are using trial-accounts you will get banned with google and AWS, too.

You're welcome