Benchmarking request for Acrobat Pro Optical Character Reader on Apple Silicon

[theorist0]

I was curious about the performance of the Optical Character Reader (OCR) in Adobe Acrobat Pro on Apple Silicon, so I created a quick benchmark and posted it on another site.  Two users reponded, one with an M1 Max and another with an M1 Ultra, and got times of 45 s and 36 s, respectively, meaning the Ultra was 45/36 => 25% faster.  However, AFAIK, the OCR is single-threaded, and doesn't use the GPU, so it's surprising there is a difference, since both devices have the same single-threaded CPU performance [Each repeated the benchmark three times and go the same time to within < 1 s.]

So I was wondering if anyone here who has an Apple Silicon Mac would be willing to repeat the benchmark and post your results here.  The benchmark instructions are below.

Please post:

Machine:
OS:
Acrobat Pro Version:
Runtime:

Here are their results:

Machine: Apple M1 Max 16" MacBook Pro
OS: MacOS 13.2.1
Acrobat Pro Version: 2022.003.20314
Runtime: 45 s

Machine: Apple Mac Studio Ultra/64GB RAM/M1/20-core CPU, 48-core GPU
OS: MacOS 13.2.1
Acrobat Pro Version: 2022.003.20314
Runtime: 36 s

Here are the steps:

[It looks complicated because I've laboriously indicated every step, but it should take under a minute to click through everything.]

0) Ensure no other significant active tasks are running.

1) In Safari (or the browser of your choice), go to this .gov website, which shows a recent Apple patent application:
https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/20230050061

2) Export it as a PDF (Safari screenshot shown):

3) Open the PDF in the lastest version of Adobe Acrobat Pro:
4) Do CMD-F and then enter “unified” in the search box, and hit “Next” (there are no instances of this, so it will be forced to search the entire document):

5) You will see this prompt. Hit “Yes”

 

6) You will see this prompt. Have your stopwatch ready. Start the watch when you hit “OK”:

7) Once the above starts runing, you will see a black progress window at the lower right:

Immediately after it completes the conversion of the last page (p 56), it will search the document and this window will appear. Stop the watch when you see it:

 

[theorist0]

Sorry, there's an issue with the USPTO web address, so I've attached the PDF  of the patent application here.

[theorist0]

....And feel free to provide benchmarks with PC's as well, since those would be interesting for comparison.

[Bernd Alheit]

The document contains text. You don't need perform OCR on this document.

[theorist0]

Ah, I see the problem.  The original document did not contain text (as evidenced by the dialog box I got after instruction #5 in first post saying "This is a scanned PDF and cannot be searched").  Unfortunately, the one I uploaded had already been converted to text, and now I can't seem to find the original.

[theorist0]

OK, I was able to find the unconverted version.  But I'm not able to attach it to a reply because it says:

[theorist0]

OK, got it working.   Apparently the problem wasn't the file type but the filename.  The original (unconverted) document is attached here.  Sorry for the inconvenience