Lofty

The setting in question controls `set_global_assignment -name ACTIVE_SERIAL_CLOCK FREQ_{100,50,25,12_5}MHZ`, which I think controls the oscillator frequency when loading bitstreams from flash. More specifically: it changes the `OPTION_REG_EXTERNAL_CLK_SPI` part of bitstream...

It seems that [AN 496](https://www.intel.com/content/www/us/en/docs/programmable/683653/current/using-the-internal-oscillator-ip-core.html) documents the Cyclone V internal oscillator at 100 MHz maximum. This is... still incredibly vague but it's better than nothing.

An interesting observation from this is that a call to `lj_carith_i64` is emitted even when `test_type` is 32-bit.

That's *one* of the issues, but the greater-scope bug is that a call is emitted in the general case, rather than inline math. Sure, the constant fold is nice, but...

@lukego Can you explain why the build failed? Is it just because I introduced a new function which changed the internal VM?

> > > hi there! :) > > note that the bitops in 5.3 and lj differs in their behavior (cuz of optimizations). I thought the only major difference was...

On Mon, 18 Feb 2019, 06:47 Luke Gorrie Here is a thought experiment too: Where will Lua, LuaJIT, and RaptorJIT be > language-wise in 10 years? > > Here's one...

So what about all the projects caught in the middle, trying to support both Lua and LuaJIT? Löve2D would be an excellent example. Maintaining compatibility with Lua is a burden,...

On Mon, 18 Feb 2019, 08:30 Luke Gorrie So what about all the projects caught in the middle, trying to support > both Lua and LuaJIT? Löve2D would be an...

Even better: we can inline [the function](https://github.com/raptorjit/raptorjit/blob/9abd48a661460904f3d169a13ce0ddafb28388a0/src/lj_carith.c#L366) pretty trivially, and if the divisor is an immedate that isn't zero or the other obscure corner case, emit an `idiv` or equivalent...