Hardware for running Golly

[NickGotts]

I'm an ignoramus when it comes to hardware. I'm planning to buy a new desktop machine (or more likely, hire my son to build one for me), as the one he built a decade ago is showing its age. Golly is probably the most demanding software I run on a regular basis - either single, highly complex patterns for billions of ticks, or Python scripts which generate and run large numbers of small patterns and test them for quiescence after a few thousand ticks. If the developers (or anyone) else can answer any or all of the following questions, I'd be most grateful:
1) As far as I can discover, there's no useful way to distribute the running of a single large pattern across multiple cores - is that still the case?
2) Is the same true for Python scripts run from within Golly?
Questions (1) and (2) are relevant to the question he asked me: is more cores, or more power per core, key?

3) Is there any useful way of getting the graphics card to handle any of the processing?

Thanks in advance,
Nick

[calcyman]

I'm an ignoramus when it comes to hardware. I'm planning to buy a new desktop machine (or more likely, hire my son to build one for me), as the one he built a decade ago is showing its age. Golly is probably the most demanding software I run on a regular basis - either single, highly complex patterns for billions of ticks, or Python scripts which generate and run large numbers of small patterns and test them for quiescence after a few thousand ticks. If the developers (or anyone) else can answer any or all of the following questions, I'd be most grateful:
1) As far as I can discover, there's no useful way to distribute the running of a single large pattern across multiple cores - is that still the case?
2) Is the same true for Python scripts run from within Golly?
Questions (1) and (2) are relevant to the question he asked me: is more cores, or more power per core, key?

3) Is there any useful way of getting the graphics card to handle any of the processing?

Thanks in advance,
Nick

For hashlife, the most important factor is the latency of the memory controller. These new AMD Ryzen CPUs tend to be very good for their combination of clock speed, cache sizes, and suchlike, so I'd recommend that.

You're correct that currently there are no parallel implementations of hashlife, but Tom and I have discussed ideas about how this may be implemented in the future.

If you're running large numbers of small patterns, that's parallelisable (because you can run different patterns on different cores). You can use GNU parallel to launch lots of programs (such as Python, or Golly+Python) concurrently.

[NickGotts]

Thanks calcyman, that's very helpful!