Let's assume that the Access procedure is running on PC1, which accesses data on the locally installed SQL Server, through the ODBC driver
Ok, now that changes things a wee bit, but again, with hardware where it stands?
Well, since the network bottleneck is gone?
Well, as long as one has a multi-core CPU. And these days, that's probably a min of 6 threads, and more CPU's?
Then that will suffice, and once again, throwing more CPU's, or more ram at this?
It's not really going to help much.
As noted, I certainly recommend having say a newer MVMe then compared to say a SSD drive.
And depending on what tools one has loaded during working?
Well, 16GIG probably the min, and say 24-32 is better. I mean, during the day, the "hogs" I have open are Visual Studio (open near all day), Outlook, and some browsers, and along with SSMS, and SQL express. These run on my computer all day long. Outlook and Visual Studio seems to be the real hogs these days. But, since I'm do lot's of web stuff all day long? Well, now I have
Visual Studio running.
SQL express server running
SSMS running.
IIS (internet services running - the web server).
Outlook
Access
And then the issue becomes do you have, or want any VM's running - I have lots of those, but I don't run them all day.
However, this conversation was limited to JUST Access, and now the question is introduction of SQL server into this mix.
I think I would trade some CPU cycles (and cores) for more disk drive speed. In many cases, if you talk to gamers etc. they will state that loading a game from a SSD is not really any slower then loading from a MVMe drive, dispite the drive being 5x faster rated.
But why?
Well, it turns out most game assets are high compressed, and it's the CPU that limiting the load times of the game (and the un-compressing and un-raveling all the game assets into memory - that's much CPU bound, not drive speed bound). So, you don't see a big jump in performance - it still much CPU bound to "assemble" the game assets.
However, in pure database land? Then the drive speed REALLY going to help more then CPU speed is.
And during the day? It's quite often that I'll take a copy of a database from say the production server, and "restore" it local.
That database is about 7-8 gigs in size. So, the faster (local) MVEe drive is VERY noticeable when I restore that database - and I often do this.
(we often do this to track down a bug - be it in Access, or from the web site - both share the same database).
And it will depend on what kind's of update(s) and SQL one is doing here. So, for me, I don't notice much speed difference in database queries etc. with say more CPU cycles. But I MOST certainly benefit and notice improved development times with faster hard drives - such as restoring and backing up databases, or working with larger projects in Visual Studio.
I thus can't share a lot more then above - but a min of 24 gigs, and a fast hard drive, and say at least 12 or more CPU threads?
After that, I really don't have first hand experience. Access is single threaded, and SQL server can use more cores/threads - but with just one user (you the developer), then even throwing more threads at SQL server not going to help, since threads are a "sideways" scaling issue.
By "sideways" scaling?
That means more users and a larger workload can be handled, but that does not mean that each single user will see their work completed faster by adding more threads/CPU's. So, a single developer box tends to not be thread starved anyway. And if you not thread starved, then adding more cores/threads not going to help you.
Going from say 3.5 ghz CPU to say 4? That will only shave 1 second off a 10 second operation - you probably not notice that.
Same for most queries and SQL stuff. However, SQL backup's and restores etc.? Oh boy, that's where I really smile and enjoy the benefits of a good fast hard drive, and such things help the developer more during a given day then say a bit more CPU cycles....
R
Albert