Model S/X deliveries with Intel-based MCU

[GatorGuy]

if you want to check if you have a new mcu or not, one good idea is to look at the software version and publish what you see. I imagine intel-based builds would differ from non-Intel, so if it's a 18.6.1 with a different hash, or 18.10 with a different hash... you will know.

This is about the least intrusive method I imagine.

I pick my car up on Thurs and will check this.

 

[Trebek1762]

We'll my SC did say a new MCU was coming, let's see if this pans out to be an actual fact

when did they say this? I'm about to reach out to my DS and see if he's heard anything...

 

[MP3Mike]

Tesla app reporting 2018.6.1 641efac. Will confirm the car shows the same hash next time I'm in it

This post shows the same build number for a S90D, so if you do indeed have an Intel based MCU build number doesn't identify it.

 

[andrewket]

I would NOT expect the build numbers or the hash to deviate across platforms. That would cause a lot of customer confusion.

The builds have the drivers and kernels for all of them.

 

[MorrisonHiker]

I would NOT expect the build numbers or the hash to deviate across platforms. That would cause a lot of customer confusion.

The builds have the drivers and kernels for all of them.

I agree. I've had various PCs over the years with Intel, AMD, etc. processors and they all had the same build of Windows, just with different drivers.

 

[DocZ]

If someone could please confirm this that would be amazing

 

[MP3Mike]

I agree. I've had various PCs over the years with Intel, AMD, etc. processors and they all had the same build of Windows, just with different drivers.

Not the same thing. Windows had different builds for different architectures: Alpha, MIPS, Intel, etc. I think now with Windows 10 the builds might be the same between x86/x64/ARM but I'm not positive.

 

[lunitiks]

New wiring diagrams for S/X were released this month. Traces of the word "MCU2"

 

[vigge50]

New wiring diagrams for S/X were released this month. Traces of the word "MCU2"

Source?

 

[tessellator]

My S75D delivered produced on 2/16 seems to still have the old MCU, at least based on the architecture reported by the browser (ArmV7)

 

[viper2ko]

My S75D delivered produced on 2/16 seems to still have the old MCU, at least based on the architecture reported by the browser (ArmV7)

Even more devastating for you. If this is true You missed it by a month

 

[bmah]

Not the same thing. Windows had different builds for different architectures: Alpha, MIPS, Intel, etc. I think now with Windows 10 the builds might be the same between x86/x64/ARM but I'm not positive.

It depends on what the hash actually represents. We were collectively hypothesizing earlier that it's a Git hash of a specific revision (because it's 7 digits and that's the default display length for Git). If so, you could certainly have builds for different CPU families that came from the same Git revision, and they'd have the same version string and hash.

If the hash is part of a checksum/HMAC of the object code / executables (say the first few digits of a SHA1 hash), then they'd be different for different CPU families.

I'm personally leaning towards the former but we're all guessing at this point.

Bruce.

 

[CSFTN]

Even more devastating for you. If this is true You missed it by a month

Devastating? Thats a pretty strong word.

 

[Shateam]

Would it be possible for someone with a recent order try to dissect their order URL to see if a new 'MCU' code has shown up? Similar to what the community revealed with Autopilot (i.e. 'APE1', 'APF1' and 'APH0' etc.)

 

[tessellator]

Devastating? Thats a pretty strong word.

Yea, if that’s the only change i’m ok. Until they start having apps that really push the MCU, having a more responsive display is not a deal breaker.

 

[viper2ko]

Yea, if that’s the only change i’m ok. Until they start having apps that really push the MCU, having a more responsive display is not a deal breaker.

The main issue for me is loading maps, pinch zoom etc

 

[verygreen]

It depends on what the hash actually represents. We were collectively hypothesizing earlier that it's a Git hash of a specific revision (because it's 7 digits and that's the default display length for Git). If so, you could certainly have builds for different CPU families that came from the same Git revision, and they'd have the same version string and hash.

If the hash is part of a checksum/HMAC of the object code / executables (say the first few digits of a SHA1 hash), then they'd be different for different CPU families.

I'm personally leaning towards the former but we're all guessing at this point.

The hash is git checkpoint, there's no question about it. You can also theoretically build multiple architectures from a single point. BUT for whatever reason Tesla does not appear to be doing it (as evident on model 3 builds), and as such I think there's a high chance the hash would be different on updated arch mcu.

Sure, it's still circumstantial, need to see a car in person to know for sure

 

[croman]

Lets not be melodramatic. Its not the biggest difference. I can't wait until the HUD drops (heh heh) or 2170 batteries with 130kwh capacity and then we'll see DEVASTATION!

 

[ReturnZero]

I also have a loaner at the moment that was brand new with 12 miles on it. It has the lighted vanity mirrors. The screen didn't really feel much different than my car, but maybe it's a little faster. If there's something I can do to find out what's in it I'm more than happy to, it's not my car so I'll pull apart whatever.

 

[arubinst]

Wouldn't it be as simple as making a quick video of the GUI in action?