• AVR Freaks

Helpful ReplyHot!Is MCHP still developing the PIC32?

Page: 12345 > Showing page 1 of 5
Author
realexander
Super Member
  • Total Posts : 231
  • Reward points : 0
  • Joined: 2006/04/08 09:50:42
  • Location: 0
  • Status: offline
2020/05/05 11:32:45 (permalink)
5 (2)

Is MCHP still developing the PIC32?

A friend of mine just pointed out that there have been no new PIC32s since 2017 (MCHP acquired Atmel in 2016). Looking at datasheets and looking for news announcement, I think he might be right.
 
If I look in the Parametric Search for Future Products, I see PIC16s, PIC18s, ATTINYs, ATSAMs, and AVRs, but no PIC32s.
 
Is the PIC32 line moribund?
 
- Bob
 
#1
JPortici
Super Member
  • Total Posts : 1091
  • Reward points : 0
  • Joined: 2012/11/17 06:27:45
  • Location: Grappaland
  • Status: online
Re: Is MCHP still developing the PIC32? 2020/05/05 12:55:49 (permalink)
5 (2)
new PIC32MZ DA have been released quite recently and new PIC32MK are going to be released (both adds ECC flash and CAN FD over the current flash and can controllers)
the previous parts will be "abandoned" in the sense that there will be probably no updates to the silicon
 
As for future moves, i heard rumors but i'm not going to report them because i don't want to jinx it. However, i expect the mips parts to fade out because mips is basically dead. One could expect RISC V core and microchip peripherals
 
ARM (cortex M) is boring. I'd like to see more accessible cortex R, otherwise i'll keep using dsPICs
post edited by JPortici - 2020/05/05 12:57:05
#2
jtemples
عُضْوٌ جَدِيد
  • Total Posts : 11891
  • Reward points : 0
  • Joined: 2004/02/13 12:31:19
  • Location: Southern California
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 13:09:16 (permalink) ☄ Helpfulby DominusT 2020/05/19 14:43:30
5 (3)
The gossip I heard from inside Microchip was that after the Atmel acquisition, PIC32 was dead, with the exception of the graphics (DA) chips.
 
#3
realexander
Super Member
  • Total Posts : 231
  • Reward points : 0
  • Joined: 2006/04/08 09:50:42
  • Location: 0
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 13:10:59 (permalink)
4.8 (5)
Thank you. It's nice to hear that there's some activity going on.
 
I don't really care about the MIPS architecture. I just want a 32 bit architecture that incorporates the Microchip peripherals and pin outs, just because I know them.
 
I'm not sure what it means for ARM to be "boring". I'd be content with an ARM core and Microchip peripherals. But I guess MCHP might prefer RISC-V to save on licensing fees.
 
- Bob
 
#4
JPortici
Super Member
  • Total Posts : 1091
  • Reward points : 0
  • Joined: 2012/11/17 06:27:45
  • Location: Grappaland
  • Status: online
Re: Is MCHP still developing the PIC32? 2020/05/05 13:22:15 (permalink)
0
Boring in the sense that i have never got the appeal of cortex M0 and M3 over say a PIC18 and a dsPIC.. really!
Every time i had to use ARM i was unimpressed with what i got in exchange for the added complexity
 
But i agree with you, i'd care less of the core as long as it had microchip peripherals and PPS as it's comes in 8/16bit parts.. PPS in PIC32 is just..
 
case in point: for CAN+USB i'll go with a PIC32MX any day. Not many offerings that aren't idiotic (stm32F1) or absurdly priced (stm32F4)
post edited by JPortici - 2020/05/05 13:24:09
#5
Stefiff
Starting Member
  • Total Posts : 84
  • Reward points : 0
  • Joined: 2012/07/15 15:26:29
  • Location: 0
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 13:25:44 (permalink)
1 (2)
Strange, but you're right. At least it seems.
There is another thing. All new versions of pic32 (MK, MM, MZ) are a major hardware-related tragedy. It is very difficult to find any module that works properly. That's why they make new revisions to existing ones. But when you look at the errata files, there are not many improvements. For now, the MZ series itself, has approximately stabilized it. Others are still very bad.
At the moment it is up-to-date are ARMs, and I guess they are trying to keep up with fashion. The cores themselves have more modifications for different applications, allowing many types of processors to be built.
 
Personal opinion.
The ARM will go down in 2 years, and will used only be in non-serious applications. Somewhere in the second half of last year, the Chinese purchased RISK-V technology and offered it for licensing, free of charge to anyone. The technology of the ARM cores has stuck to its limit, there is nowhere else to improve. This is understood by all the bigs. Since the end of last year, virtually all major players have officially stated that they are stopping any investment in the development / using of ARM cores. All officially will puts all resources into developing RISK-V, and their future processors will be with RISK-V cores. Practically, wherever the ARM has stopped, RISK-V begins to develop from there.
The main problem with embedding in consumer electronics was the lack of a video processor. But the Chinese have introduced, by memory, a second-generation video processor that is similar in scope to the current used in ARM (Mali). It is expected to launch a third-generation GPU in the near future, and then they will flood the market with any modifications of the chips. Just have all set up very seriously, and pour very large amounts. Soon there will be visible results.
#6
JPortici
Super Member
  • Total Posts : 1091
  • Reward points : 0
  • Joined: 2012/11/17 06:27:45
  • Location: Grappaland
  • Status: online
Re: Is MCHP still developing the PIC32? 2020/05/05 13:29:07 (permalink)
4.5 (2)
missed chances:
-PIC32MM, the ADC and PWM erratas render it useless over a PIC32MX and it never was fixed, shame because i really liked the part per se in the projects i used it.
-PIC32MK, the amount of erratas that impact actual performance is just too much. Pity. There are not many parts with 2 USB ports in a microcontroller with also all of that PWM and ADC stuff. Not many use cases for 2 USB but it's by far the cheapest part around by 2-3 euros at least. Still have hope for future silicons as above
 
#7
realexander
Super Member
  • Total Posts : 231
  • Reward points : 0
  • Joined: 2006/04/08 09:50:42
  • Location: 0
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 13:34:28 (permalink)
4.83 (6)
Stefiff
Somewhere in the second half of last year, the Chinese purchased RISK-V technology and offered it for licensing, free of charge to anyone.

 
RISC-V is not owned by "the Chinese". Check out the Wikipedia article on it.
 
- Bob
 
#8
dlindbergh
Super Member
  • Total Posts : 481
  • Reward points : 0
  • Joined: 2006/09/11 18:17:11
  • Location: Boston USA
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 13:35:56 (permalink)
5 (1)
StefiffAll new versions of pic32 (MK, MM, MZ) are a major hardware-related tragedy.

 
I've been under the impression that the MZ EF parts (not the original MZ) mostly work as planned - at least, the errata list is short. No?
 
I haven't tried them yet, still using PIC32MX. But a 12 bit ADC and FPU would be nice sometimes.
#9
NKurzman
A Guy on the Net
  • Total Posts : 18786
  • Reward points : 0
  • Joined: 2008/01/16 19:33:48
  • Location: 0
  • Status: online
Re: Is MCHP still developing the PIC32? 2020/05/05 13:36:51 (permalink)
4 (2)
I used a pic32mx530. Uarts and a/d worked fine.
Pic32mz EFH.
UARTs, spi, usb, Ethernet, I2c, bootloader.
All work well. Now the ECH, epic fail.
#10
Stefiff
Starting Member
  • Total Posts : 84
  • Reward points : 0
  • Joined: 2012/07/15 15:26:29
  • Location: 0
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 14:27:03 (permalink)
2 (3)
realexander
Stefiff
Somewhere in the second half of last year, the Chinese purchased RISK-V technology and offered it for licensing, free of charge to anyone.

 
RISC-V is not owned by "the Chinese". Check out the Wikipedia article on it.
 
- Bob
 

 
About RISC-V Foundation.
It all started with Trump. The Chinese have taken the threat of limiting the use of ARM cores by the Chinese very seriously. Then they were searching for a quick decision. The RISC-V was selling itself because it was financially bad. They bought them. Then, due to some legal reasons,  they were separated it into an independent company that would fictionally develop on its own. Everything evolved within a month. I remember it very well, because there are articles with all the changes that happened, literally in every 4-5 days. Trump cannot impose any restrictions on them now.
That's why they offered the licenses for free. In order to attract more to their side and invest more resources in RISC-V development.
#11
realexander
Super Member
  • Total Posts : 231
  • Reward points : 0
  • Joined: 2006/04/08 09:50:42
  • Location: 0
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 15:00:19 (permalink)
5 (2)
RISC-V had an open license before Trump came along. It's true that Chinese companies want to use RISC-V and that it moved to Switzerland to avoid trade issues.
 
Can you cite a source for "the Chinese" "buying" RISC-V?
 
#12
Stefiff
Starting Member
  • Total Posts : 84
  • Reward points : 0
  • Joined: 2012/07/15 15:26:29
  • Location: 0
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 15:27:34 (permalink)
1 (2)
realexander
Can you cite a source for "the Chinese" "buying" RISC-V?

 
I'm sorry but I can't provide a link to this information. It was all a long time ago. It was very strange to me because so many changes had not happened on such a global scale in such a short time. Literally, indeed, in 4-5 days, there was some change. It usually takes months to change something in this industry. And here it was very fast.
============
The important thing is to expect new nice processors soon!
 
Perhaps the most shocking thing for me was reading several ARM processors files in the 400-600MHz range. The consumption at 25C is 0.6-0.7A, at 80C - 3A. This is not a normal processor for me.
 
#13
dlindbergh
Super Member
  • Total Posts : 481
  • Reward points : 0
  • Joined: 2006/09/11 18:17:11
  • Location: Boston USA
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/05 15:48:13 (permalink)
4.75 (4)
Stefiff It all started with Trump. The Chinese have taken the threat of limiting the use of ARM cores by the Chinese very seriously.

 
I think you may be confusing MIPS with RISC-V. At one point I think the Chinese did buy the rights (or company that owned the rights) to the MIPS architecture, and were supposedly going to make it the "Chinese national CPU architecture" (and supposedly kill off ARM and Intel).
 
There was going to be the "Loongson" (literally "Dragon Chip") family of CPUs: https://en.wikipedia.org/wiki/Loongson
 
I don't think all of that went much of anywhere - the interest has shifted to RISC-V. Much like what happened with the Chinese attempt to make a Chineseified Linux a "national operating system" - nothing. Industrial policy is easy to make press releases about, but hard to make happen in the real world.
 
Per https://en.wikipedia.org/wiki/RISC-V#History , RISC-V was under a BSD license (it started at UC Berkeley - funded by Microsoft and....Intel!) and since has switched to a Creative Commons license. I don't think China (or anybody else) can own such a thing - it has been open from the beginning. (But yes, the Chinese are among those heavily investing in RISC-V now - but so are Google, Nvidia, Samsung, and others.)
#14
jdeguire
Super Member
  • Total Posts : 565
  • Reward points : 0
  • Joined: 2012/01/13 07:48:44
  • Location: United States
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/06 11:23:58 (permalink)
5 (2)
Microchip is working on a PIC32C series, which will have Cortex-M CPUs--M4 for the PIC32CX and M7 for the PIC32CZ.  There isn't much known about them yet other than you can find files for them in Harmony 2.04, but it is possible that they'll have Microchip as opposed to Atmel peripherals.
 
I went to MASTERs last year and my impression is that there was a lot of focus on the SAM parts.  A couple classes I took used the SAM parts and never mentioned the PIC32, for example.  That said, Microchip will certainly keep supporting the current MIPS parts if you want to use those, even if they don't introduce new ones.  I suspect part of the heavy focus on the SAM parts is that they are new to the Microchip crowd rather than an attempt at downplaying their place in the lineup.
 
Harmony 3 release notes make reference to the "PIC32MZ W1", so there is a new high-end PIC32 device.
 
In case anyone is curious about how the MIPS and Arm cores compare:
Some time ago, I got CoreMark running on a test project and used that to compare a few devices.  You can download the CoreMark benchmark from GitHub if you want to do something similar.  I used just our basic compiler options we use for most projects ("-O1" mainly) and whatever version of Harmony 3 was available when I did this last summer.  From those tests, it would seem that the MIPS M4K core is slightly faster per MHz than the Cortex-M4 (PIC32MX795 vs SAME54).  The E54 got about a 35% higher score with 50% higher clock speed, so the difference isn't huge.  The Cortex-M7 outperforms the MIPS Warrior M-5150 core (SAME70 vs. PIC32MZ EF) on a per-clock basis, however, scoring a bit over 70% higher with a 50% higher clock speed.  Note that the regular CoreMark benchmark is integer-only and that these results were taken with whatever default values CoreMark lets you use.  Also, I make no claims that my test is perfect and may be flawed in some way, so take this with a grain of salt.
 
There is a CoreMark PRO available that is supposed to be more extensive, but I haven't had a chance to set that up yet.
#15
NKurzman
A Guy on the Net
  • Total Posts : 18786
  • Reward points : 0
  • Joined: 2008/01/16 19:33:48
  • Location: 0
  • Status: online
Re: Is MCHP still developing the PIC32? 2020/05/06 11:37:34 (permalink)
5 (3)
You may Need to look a little deep to get a Good Comparison.  The Overhead an latency and the I/O Speed are thing that affect embedded performance too.
#16
jdeguire
Super Member
  • Total Posts : 565
  • Reward points : 0
  • Joined: 2012/01/13 07:48:44
  • Location: United States
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/06 13:28:37 (permalink)
0
Indeed, and I suspect that the flash on the PIC32 devices will help them along.  From what I recall, the flash operates at about 30MHz on the PIC32MX795, 66MHz on the PIC32MZ, and ~24MHz on the SAME54 and E70.  All of them use 128-bit busses.  This means that on purely sequential code, the flash will be a HUGE bottleneck on the SAM E70 @ 300MHz, but the flash on the other devices should be able to keep up.  At least, that's my semi-educated guess.
 
From what I remember, the CoreMark test uses a bunch of small loops to run and so the cache can save the E70.  A test I want to try in the future is to turn off the instruction caches on the MZ and E70 and re-run CoreMark.  My thought is that this might be more like what the processors will have to deal with in purely sequential code when the cache can't save them.
 
I had a separate test that tried to test interrupt latency by manually forcing a interrupt to occur and then setting/clearing an IO pin as the processor entered and exited the interrupt.  I don't remember the exact numbers, but the Arm Cortex-M parts did better than than MIPS parts.  The Cortex-M cores have register auto-stacking and unstacking features built in to help with latency and have half the registers to save vs. the MIPS cores.  EDIT:  I did use the shadow registers on the PIC32MZ and the Arm parts were still better.
post edited by jdeguire - 2020/05/06 13:33:31
#17
simong123
Lab Member No. 003
  • Total Posts : 1391
  • Reward points : 0
  • Joined: 2012/02/07 18:21:03
  • Location: Future Gadget Lab (UK Branch)
  • Status: online
Re: Is MCHP still developing the PIC32? 2020/05/06 16:19:49 (permalink)
5 (5)
jdeguire
I had a separate test that tried to test interrupt latency by manually forcing a interrupt to occur and then setting/clearing an IO pin as the processor entered and exited the interrupt.  I don't remember the exact numbers, but the Arm Cortex-M parts did better than than MIPS parts.  The Cortex-M cores have register auto-stacking and unstacking features built in to help with latency and have half the registers to save vs. the MIPS cores.  EDIT:  I did use the shadow registers on the PIC32MZ and the Arm parts were still better.

The 'MZ also has auto prologue/epilogue and vector pre-fetch which speed up interrupts greatly. However XC32 doesn't support it natively, so it requires a little messing to get it to work. It reduces latency to ~11clks @200MHz.
You have to enable the feature in CP0 and declare the interrupt functions slightly differently e.g.
void __attribute__((at_vector(_CHANGE_NOTICE_D_VECTOR),naked,nomips16)) CNDInterrupt(void)
{
    .....

    asm volatile("iret");
}

Note this is also all-or-nothing, all interrupts must use the extensions, and assumes shadow sets enabled.
 
Edit: Added iret to interupt function.
post edited by simong123 - 2020/05/06 18:46:24
#18
Stefiff
Starting Member
  • Total Posts : 84
  • Reward points : 0
  • Joined: 2012/07/15 15:26:29
  • Location: 0
  • Status: offline
Re: Is MCHP still developing the PIC32? 2020/05/07 01:25:08 (permalink)
0
The tests usually use the CPU, arithmetic processor and RAM. The PIC32 has wait cycles, the ARM has a small RAM with the same frequency that the CPU runs. In most cases, it is clear who will do better.
 
With predominantly mathematical operations, the ARM I suppose will be better.
More important to me is the speed at which the information is transferred between the modules, and from the outside to the inner/main RAM through a module for a long time.
In ARM we have some small RAM that is connected to certain modules. BUT then it should be transferred from there to the main RAM. And in this whole process, outside to the main RAM, caching no longer does much work. Even if the peripheral module itself works quickly, locally, until everything passes through the CPU, and is written to the main RAM, in this case the different architecture is shows up.
In this “PIC32MK - Stupid UART/DMA/CPU Design - Be Careful!!” topic I wrote:
Someone, somewhere, wrote that an ARM at 500MHz equals to PIC32 at 200MHz for gpio.
 
That's exactly what it was about, not for math tests.
In memory, they had taken a simple test then. They had cycled CPU on one gpio to switch it cyclically, and with an oscilloscope measured what frequency could be obtained on that pin.
I am far from entering into any disputes. What I saw, I wrote.
 
For example, last week I struggled with the SDHC controller of PIC32MZ2064DAS176. Apart from the fact that there is some kind of hardware bug, the real maximum speed that is obtained when reading a card and writing in the internal DDR2 of 1MB is about 11-12MB / s. When stored in the internal RAM, the speed rises slightly to 14-15MB / s.
Which is quite far from the theoretical 25MB / s. The SDHC module itself clocks the card at 50MHz. When using Harmony 3 it is even slower. For me, this is a real test of the capabilities of the PIC32MZ.
#19
JPortici
Super Member
  • Total Posts : 1091
  • Reward points : 0
  • Joined: 2012/11/17 06:27:45
  • Location: Grappaland
  • Status: online
Re: Is MCHP still developing the PIC32? 2020/05/07 02:12:02 (permalink)
4.5 (2)
Stefiff
The tests usually use the CPU, arithmetic processor and RAM. The PIC32 has wait cycles, the ARM has a small RAM with the same frequency that the CPU runs. In most cases, it is clear who will do better.

I was under the impression that you could also run code from RAM in the PIC32MZ, and not just a tiny area of TCM
what am i missing?
 
#20
Page: 12345 > Showing page 1 of 5
Jump to:
© 2020 APG vNext Commercial Version 4.5