Somewhat inspired by somebody's question on Reddit, I thought it might be useful to do a little infodump on what I know of how graphing speed differs between 8x calculator models and OS versions.
The main conclusion to draw here is that asymptote detection is very slow, so if your graphing performance is not as good as you'd expect you may be able to turn it off and make graphing much faster (at the cost of vertical asymptotes being drawn as vertical lines rather than discontinuities).
Rather annoyingly, the Internet Archive doesn't seem to have a copy of the threads on the Detached Solutions forum linked from the OS 2.30 discussion on calcg, and the DS forum seems to have been removed at some point (rather reasonably, since it was basically dead and probably tedious to handle spam). I'm following up with the DS staff I have contacts for to see if anybody kept a copy.
- The baseline TI-83+ is the baseline. I have nothing interesting to say about that.
- The 83+ SE runs the same OS versions as the baseline 83+ but runs the CPU at 15 MHz rather than the baseline 8 MHz, so should be about twice as fast.
- The monochrome 84 models (TI-84+ and TI-84+ SE) run the CPU at the same 15 MHz as the 83+ SE, so for most operations they should also be about twice as fast as the baseline 83+.
However, OS 2.30 added asymptote detection when graphing functions which makes graphing much slower than the baseline. Sadly, versions before 2.30 (2.21, 2.22) are said to have various bugs that make them problematic for various use, and later versions with TestGuard (2.40 and up) or MathPrint (2.53MP and up) might be required for some uses.
Graphing in Parametric mode doesn't support asymptote detection so should be as fast as would be expected based on the CPU speed.
- The TI-84 Plus Color Silver Edition (CSE) is very slow all around because its high-resolution color screen is a poor fit for its architecture which is largely unchanged from the monochrome calculators. It does however have an option to disable asymptote detection in the FORMAT (2ND->ZOOM) menu, so disabling that should regain some graphing speed.
- The TI-84 Plus CE (and all variants such as the TI-83 Premium CE and various Python Editions) has basically the same feature set as the CSE but is generally fast- exactly how much so I'm not sure since I don't own one myself nor have I tried to do any true comparison.
The main conclusion to draw here is that asymptote detection is very slow, so if your graphing performance is not as good as you'd expect you may be able to turn it off and make graphing much faster (at the cost of vertical asymptotes being drawn as vertical lines rather than discontinuities).
Rather annoyingly, the Internet Archive doesn't seem to have a copy of the threads on the Detached Solutions forum linked from the OS 2.30 discussion on calcg, and the DS forum seems to have been removed at some point (rather reasonably, since it was basically dead and probably tedious to handle spam). I'm following up with the DS staff I have contacts for to see if anybody kept a copy.