Optuma to run on Arm Based Computers ie Apple M1 or Windows Arm64

I actually run a M2 MacBook Pro myself now.

Optuma runs well but one setting I had to change was to stop Optuma from downloading data in parallel, which seemed to make it very slow while it was updating data. See the image below. This will be on by default.

As far as RAM is concerned, 16GB is the minimum.

All the best

Mathew
Screenshot 2024-03-25 at 8.40.14 am

1 Like

Hi Thomas,

I own a MacBook Pro M2 Max with 32GB Ram running Parallels 19 Pro Edition.

If you wish to pump up the RAM, you’ll need the Pro not Standard edition. This will then allow you to allocate up to 128.

From my personal testing, I found a good bump in speed going from 8 to 16, but when I tried to 24 I found the time didn’t change much.

If you are using generally though, the speed won’t vary too much I was purely testing to see the difference.

If I open the ASX200 then using arrow keys to scroll through front to back, 8GB was 63s, 16 and 24 was 45s.

As Mathew mentioned runs really well, and I wish I took the 64GB ram option, but even with 32GB and heaps of apps open (currently 15 open), and Parallels set with 10 Processor and 24GB I find everything works exceptionally well. Less than 1GB being used in Swap.

It’s a beast and while I haven’t tested large watchlists like you’ve mentioned as a speed test comparison for you, to open ASX 200 under symbol list just default chart, to load is 11sec and to return screen ready to go another 7sec, so in total was only 18sec. If you try the same on your Windows machine might give you an insight to a speed comparison.

Mathew, thanks for the tip on that setting, will give it a try.

All the best with your new machine!!

Cheers
Scott

1 Like

Hi Mathew and Scott,

Thank you very much for your efforts.

I think I will try the Mac Studio when hopefully a new version with the M3 chip is available in the summer.

I think I will set it up with 64 GB RAM. Better safe than sorry. The RAM issue is becoming more and more of a bottleneck.

Thanks,
Thomas

Hi Thomas,

As a matter of interest, what’s the time comparison on your current windows machine to open ASX 200 to be fully loaded using default simple chart?

Not sure how familiar you are with Macs, but their unified memory is a lot more efficient that Windows machines from everything I’ve read and experienced.

M3 is enticing, but dollar / value, I always grab the refurbished models, this time around I saved over $1500 at purchase, for virtually a new machine.

All the best, cheers
Scott

Hi Optuma Team,

Perhaps you can help me answer Scott’s question about the time it takes for a Windows computer to open the ASX 200 to fully load with the default simple chart.

As I have not subscribed to the Australian database, I cannot test it.

Thanks,
Thomas

Don’t hate me but it is much faster on my Windows PC both in Optuma and the raw Optuma engine. The new Quant work I am doing processes an S&P 500 model in around 23 seconds on the Mac and 16 seconds on Windows.

I think Apple makes some great laptops etc. I chose the Air when they came out in 2010(ish) because it was so great to travel with. I got the M2 Pro because I’m keen to see how it works now. We also have an M2 Mac Mini for Video editing.

The benefit that Apple has right now with the M Series is that they achieve better speed with similar specs (due to the reduced instruction set of the CPU). The massive disadvantage is that you can not upgrade the RAM and the prices they charge for the similar RAM and storage is crazy.

But today… unless you are doing video editing a Microsoft Surface laptop or a Dell XPS desktop is a better solution if Optuma is your primary application.

The point is that if you buy a Mac because you want a Mac then Optuma will run fine on it with Parallels. But don’t buy it because you want Optuma to run better. Also, after many years of using PCs, prepare for frustration as keys are different and the muscle memory to CTRL-C takes a long time to change.

I want to love the Macs, they are safer to use in regard to viruses and phishing scams, and I love the integration with other Apple products, but there are compromises you have to be prepared for.

1 Like

Hi Mathew,

Thank you for your efforts and the time you took to run the performance tests.

I think I will stick with a Windows PC when I buy a new one this year. For the Quant engine, the speed at which the tests are run is of great importance.

Thanks a lot!
Thomas

Hi Mathew,

Thank you for taking the time to do a comparison, greatly appreciated, by the looks of it that’s a descent speed difference if using Optuma is your main work application.

What are your thoughts moving forward should Windows do a native version of ARM once Delphi supports ARM also. Would you expect the speed performance to improve, currently Parallels 19 does a great job but it’s still not running it natively. You educated me on a post back in 2021 about future timeframes, how’s it all tracking, can you share any update?

As I often use my laptop on battery mode when travelling, battery life is a massive game changer on these new machines, so I’ll stick with losing a few seconds of performance vs running a windows laptop, I’m assuming the time of 23 sec was on the desktop?

While not being able to upgrade memory I know is sore point for many users, the cost of life I believe cannot be compared to, time is money after all, and I’m sure your M2 MacPro will run just as fast in three or six years from now, just like your old MacBook Air.

The muscle memory isn’t too hard to get used to, I’ve used both Mac and PC fairly equally over 30 years and something that we easily adjust to.

Keep up the great work, I’m only starting to get into the using Optuma’s scanning ie Signal & Trade Tester but with my limited skills have been really impressed. Back to learning more coding.

AI for trading morning forward… Interesting times ahead!

Thanks again!

Cheers
Scott

There are a couple of things that need to happen for us to deliver Optuma on a Mac or Windows Arm.

  1. Delphi needs to support it.
  2. Our code needs to be able to cross compile.

The time to do that needs to be balanced with our other developments.

The primary focus for us is to get Optuma in the browser. Then platforms do not matter. All the processing will happen on our servers. This is why I am working at shaving milli-seconds off scans and tests.

The core of the project is an engine that does all the heavy lifting. It’s based on an elastic model that can ramp up extra servers as needed.

In front of that we have a web application that will be very similar to Optuma now expecpt that I’m making it simpler to use. I’ve watched where people struggle with the features and I’m trying to improve that. I’m aiming to address 80% of the use cases in the web app. For the reamining 20%, there is always the windows application which we will continue to maintain.

As resources allow, we’ll add Mobile Apps and also Optuma 3 that uses the same engine, either locally or on the server.

I assure you that the time taken for all of this is frustrating me more than anyone else, but I have to make sure the foundation is rock solid because so much will be built on top of it.

I’m committed to the first beta being in place by Q4 no matter what. I can’t aford to let it go on any longer. I’ll be asking for feedback from people before hand so if you have not already signed up, go to app.optuma.com and register.