Studio6 slower compile time

Go To Last Post
139 posts / 0 new

Pages

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You are correct... give me the simple one!

In fact, I am just frazzled ever since AS6.

I don't really enjoy C programming, but did some for a project I was working on. I was using CodeVision and found it to be a perfect work of art. Then it stopped being able to talk to my programmer as I upgraded to an AVRISP-II and AS6 due to needing to use the XMega256 and XMega384.

So I tried to use GCC. It gave me a similar bad taste as this whole GAS nightmare - as if all the larger AVRs were just kind of half-assed supported. 32k limits, progmem workarounds, 3rd party macros... ugh! I actually got so disgusted one night I deleted all of my project files and gave up on it completely.

Now I am getting there again. This AS6 compile bug was annoying, but the sinking feeling I am getting after 24 hours of working with GAS and looking at others examples is coming back to haunt me again. Once again, it seems like using the wrong sized wrench on a rusty stripped bolt.

That's what is worrying me. I do this all for fun, and if GAS is going to be Atmel's future answer to assembly, then I am gone. I still enjoy working with CPLDs and FPGAs because you can get right in there and create gates. You don't have to fight the toolchain or use half-baked workarounds to program them.

I am going to put this all away for a few days now because I have really had it.

When I calm down and come back, I will be searching for the answer to the only question I have after this experience...

What is the future for Atmel Assembly?
Is it going to be their own assembler (yay!)
Or will they drop it for GAS (Thhhpppph!)

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I was using CodeVision and found it to be a perfect work of art.
And there is rumblings that CV is being integrated into AS6 finally, so it will either help as one is not forced to use the "toolchain" or break CV. :?

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Nice!

I know I will take hell for this next comment, but GCC should implicitly state that it is not compatible with XMega, or really any AVR with large program memory or even the ramp-Z register.

I would go as far as to take out the option the even select the Xmega from it altogether.

Atmel must have a large enough budget to roll out proper tools that fully support their amazing silicon, no?

A good example to follow would be Xilinx. When the IDE supports something new like Spartan6, then it really supports it all the way, not half-assed with a few macros written by hackers.

I know, GCC grew into the Xmega, not the other way around, but why should I care? I am the user, and I want perfection. It's not too much to ask for, I have seen many amazing IDEs and toolchains in my journeys.

Same goes for this GAS deal. Why try to force these amazing uCs into some general use tool that has been around since vacuum tubes were all the rage?

How much would it really cost the company to hire a small team to write their own C compiler and up to date assembler specific to only Atmel chips? Hell, I could probably even fund the damn thing with my own R&D budget.

Ok, I should have stayed away for a while longer... getting hotter than a 1 ohm resisor in a 1970's TV.

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AtomicZombie wrote:
How much would it really cost the company to hire a small team to write their own C compiler
God forbid they listen to you and develop something Microsoft-based.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Microsoft? Didn't they do that windows thing? You mean hardware programmers actually use windows?

... crazy talk.

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

How the heck does the rjmp to "TIMER0_OVF_vect:" happen here? I know the compiler is doing this, but if that's the case,

I already explained that in this thread?!?

The vector jumps come as part of the CRT. It's your choice as to whether you use that mechanism or use -nostartfiles in which case YOUR first linked file positions at 0x0000 and you can .org to the vector locations just as in Atmel assembler.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Had a night to reflect.
Normally, I don't get frustrated with things, but this really pissed me off, so I calmed down to to figure out why.

I guess my ONLY reason for being an AVR fanboy is because it has an amazing instruction set that allows hacker like me to do incredible things with the hardware. My latest project pumps the XMega up to 57MHz and does 4 channel sample sound, and full chroma/luma synthesis right from software. This is FPGA territory folks!

You cannot do this in assembly with any other 8 bitter. Not a chance in hell. And you can't get very far with an ARM in assembly - and I have tried many times. It's an ugly scary place.

So when I finally hit a point in my hobby when I needed C (doing a bunch of 3D stuff on XMega), I gave GCC a whirl. Sadly, I realize that the toolchain is not ready for the XMega. It just isn't suited for such a large 8 bit / pseudo 24 bit chip.

Then AS6 farked up the Atmel assembly compiler for using XMegas. So I had a quick battle with GAS.

Once again, I got that feeling that this offering was far from ready for the XMega, or even pure assembly language for that matter.

So the only reason I like AVRs to begin with was slipping away. being nothing more than a part time hacker / fanboy, the good times seem to be over.

I woke up and realized...

If I am going to have to migrate to C because every toolchain offering is soon going to leave assembly freaks behind, then why the hell would I even bother with the XMega?

I can get 10x the bang for 1/2 the price in an ARM and use a high level toolchain that is more mature.

I mean, why would anyone use an XMega anyhow? Simple : great instruction set and huge program memory right?

WELL THESE ARE THE THINGS LEAST SUPPORTED IN GNU!

So moderator, please close this thread - I have extracted what I need from it and appreciate all the help everyone offered.

In the end, I was really just searching for a personal answer - why would I use Xmega?

Cheers!
Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hey AtomicZombie!

Did you see Deans post above where he says that if he collapses one of the folders in the Solution Explorere he oes not get the ridiculously long build times?

Also, since you said above that all you want is a straight forward editor and a way to assemble, then have you contemplated using some good free editor (e.g. Notepad++) and running the Atmel assembler as a comand from within that? (No Atmel Studio with lots of stuff that re-renders itself again and again during a long build. As people have stated above: The sluggishness you experienced to start with was in the IDE - not in Atmels Assembler as such.)

As of January 15, 2018, Site fix-up work has begun! Now do your part and report any bugs or deficiencies here

No guarantees, but if we don't report problems they won't get much of  a chance to be fixed! Details/discussions at link given just above.

 

"Some questions have no answers."[C Baird] "There comes a point where the spoon-feeding has to stop and the independent thinking has to start." [C Lawson] "There are always ways to disagree, without being disagreeable."[E Weddington] "Words represent concepts. Use the wrong words, communicate the wrong concept." [J Morin] "Persistence only goes so far if you set yourself up for failure." [Kartman]

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes, I have found harmony with using AS6 and the assembler thanks to the help provided here.

My fear though is for the future of the Atmel Assembler. Are they offering GAS as an alternative or because of a future migration?

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

My fear though is for the future of the Atmel Assembler. Are they offering GAS as an alternative or because of a future migration?

Both have been offered for many, many years now in tandem, and there's no plans that I know of to yank it out from under you. In any case if all you need is an assembler for a device you already have, even if it was hypothetically removed in the future - why would it matter? You could just use the tools you have today that work.

After all, with a bare assembler the only thing you'll be missing is new part support.

- Dean :twisted:

Make Atmel Studio better with my free extensions. Open source and feedback welcome!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

Are they offering GAS as an alternative or because of a future migration?

No gas and the Atmel assembler have always lived side by side as long as each has existed. The only thing that's happened recently is that Atmel wanted to embed a free C compiler in their "Studio" tool and the obvious (only?) choice is avr-gcc. It just so happens that the way that C compiler (and C++ compiler) work is to generate Asm source and there's a matched assembler (avr-as) that knows how to assemble that source. Almost as a side effect it means you can also use that same assembler in isolation if you want. I don't think there's any chance one wil replace the other. There are far too many legacy products and far too much Asm library code written for the Atmel dialect for them to ever ditch it.

also the situation with avr-as .h support is a bit ridiculous. Someone at Atmel made a concious decision to support all the bit masks as enums rather than #defines. It was the wrong decision. Maybe they'll correct it one day.

As for the expression evaluator of GNU as. It's different to Atmel so there may not be a one to one correspondence in how things are done. In fact often the easiest way to find the correct as syntax to use is to express what you want to have happen in a few lines of C and see what as source code the C compiler generates.

If you move to larger micros where GCC is used more widely you will find that most startup code is written in GNU as syntax but beyond that you are write, no one writes large projects in Asm on such micros. There's often 100's if not 1,000's of KB of code involved and you just cannot produce that much code in Asm cost effectively. A lot (most?) of the as syntax is for the benefit of the compiler writers so they can convey high level language constructs effectively to the assembler.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'll file an enhancement request to the toolchain header team to add in additional defines for the group configs when in assembly mode, but I fear it won't be prioritized.

- Dean :twisted:

Make Atmel Studio better with my free extensions. Open source and feedback welcome!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok, thanks again for the patience and help!

I just overdid it and blew a gasket. When I do these projects,I usually go for 2 days without sleep until I win. I was over tired, and ended up running in circles this time.

It's all good again.

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

I'll file an enhancement request to the toolchain header team to add in additional defines for the group configs when in assembly mode, but I fear it won't be prioritized.

Dean,

You saw my utility above? You are welcome to it. It converts the enums to #defines suitable for Asm use in about 2-3 seconds per file.

Cliff

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'll direct the team to it - however, since the existing header files are automatically generated from a set of master header files, it would be better to adjust the main conversion tool rather than adding a secondary post-processing stage.

- Dean :twisted:

Make Atmel Studio better with my free extensions. Open source and feedback welcome!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

it would be better to adjust the main conversion tool rather than adding a secondary post-processing stage.


True enough - the above would just be a short term fix that could be given to anyone who asks when using the current software. Or just give them the set of iox*-asm.h files it generates.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

I'll direct the team to it - however, since the existing header files are automatically generated from a set of master header files, it would be better to adjust the main conversion tool rather than adding a secondary post-processing stage.

Do you want to add the request? I do concur that I don't see this becoming a priority at the current time.

Regarding the atmel assembler, I see no reason why it should be discontinued. There are at least still commits comming into it...

All header files are generated based internal master files, so there will/should never be post proccessing to them (there have in the past been to much manual adjusts on the files so that maintenance were the literate nightmare)... However, all improvement requests that we get will be stored and probably dealt with in some way (tm) in time (tm)..

:: Morten

 

(yes, I work for Atmel, yes, I do this in my spare time, now stop sending PMs)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

Do you want to add the request? I do concur that I don't see this becoming a priority at the current time.

Already added, can't remember the ID but do a Jira search on my profile in the Device Headers project. I agree it's probably not going to be fixed any time soon, but it doesn't hurt to add a request anyway.

Can you reproduce the bug the OP has? I updated Jira with my latest findings - it's got the sample application that triggers the fault attached to it.

- Dean :twisted:

Make Atmel Studio better with my free extensions. Open source and feedback welcome!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Time taken to update labels at solution explorer is fixed and building the Assembler projects with Atmel Studio 6.1 beta. You can find the AS 6.1 beta at http://www.atmel.com/tools/atmel...

Regards,
Deena

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The annoying slow build/compile time is back to haunt me after six years!

 

Win-10, Studio 7.0.1645

 

Seems to be about a 50-50 chance of either one of these when hitting "Build" or "Start Without Debugging"...

 

1) Instant build and programming as expected for the last six years.

2) 15-30 seconds of waiting with "Build started..." and animated icon running.

 

It is super annoying on a project I am doing right now since I am programming many times to see instant results.

It does not happen on all of the computers I use, jsut a few.

 

It's not the project or files, as it happens even with a blank fresh project.

Also tried USB drive, local drive, network drive... same random problem.

 

Any ideas?

Brad

 

I Like to Build Stuff : http://www.AtomicZombie.com

Last Edited: Mon. Nov 26, 2018 - 09:01 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Does the title bar say something like "(unresponsive)" after the "Atmel Studio" text? Do you get the spinning cursor?

 

I continually have this problem with AS7 on an HP laptop + Win10 with seemingly "enough" memory. It seems like something is cached in virtual disk and it takes forever to get it out. It usually does not say "unresponsive" in this case but always do get the spinning cursor.

 

Also have this problem, almost 100% of the time, if Studio (and the machine) have been idle for any length of time, say 15-20 minutes or more. Here, the title bar often (but not always) says "unresponsive".

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Last Edited: Mon. Nov 26, 2018 - 09:14 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Nope, it just runs as if I am compiling 20 million lines of code!

Here is the odd thing, if I keep on hitting build, it will build 4-6 more times fast and then do the slow down for one build.

 

I have just downloaded the next AS7 build, and will try that now.

 

Thanks,

Brad

 

ka7ehk wrote:

Does the title bar say something like "(unresponsive)" after the "Atmel Studio" text? Do you get the spinning cursor?

 

I continually have this problem with AS7 on an HP laptop + Win10 with seemingly "enough" memory. It seems like something is cached in virtual disk and it takes forever to get it out. It usually does not say "unresponsive" in this case but always do get the spinning cursor.

 

Also have this problem, almost 100% of the time, if Studio (and the machine) have been idle for any length of time, say 15-20 minutes or more. Here, the title bar often (but not always) says "unresponsive".

 

Jim

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok, just installed 7.0.1931, and no change.

 

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So, what happens on your machine the one-in-six times it's slow?

:: Morten

 

(yes, I work for Atmel, yes, I do this in my spare time, now stop sending PMs)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I get the vague feeling that it is, somehow, the OS. Something about virtual disks, maybe exceeding the available RAM space, then being forced onto the real disk. Just a hunch.

 

But, clearly, it is something special about Studio because other apps running at the same time (e.g. TeamViewer) do not experience this. I do have a lot of "data" in the project, about 30 .c files and similar number of .h.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Last Edited: Mon. Nov 26, 2018 - 09:37 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Virus scanner,?

#1 This forum helps those that help themselves

#2 All grounds are not created equal

#3 How have you proved that your chip is running at xxMHz?

#4 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand." - Heater's ex-boss

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Had forgotten about virus scanner here. Need to check that.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Looked at task manager, and nothing is using much memory or cpu, even when Studio takes that mini-vacation.

 

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AtomicZombie wrote:

...

2) 15-30 seconds of waiting with "Build started..." and animated icon running.

...

It does not happen on all of the computers I use, jsut a few.

 

Hmm. that is a very long time...

What is different on the PCs ?

Maybe it is phoning home - are those slower computers internet connected ?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
I get the vague feeling that it is, somehow, the OS.
Others have felt that before and sure enough :

https://www.avrfreaks.net/forum/slow-compiling-win10-update-issue-maybe#comment-2467501

Windows 10 1809 is in a slow roll-out.

 


https://www.avrfreaks.net/forum/another-win10-rant-frustration-part-2?page=2#comment-2467891

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks, that seems to have fixed the issue so far.

Went from win10 1803.17134.1 to 1803.17134.48

 

My invoice for 7 wasted hours over the course of the week has been mailed to Microsoft.

I sure hope they pay me soon. I have been billing them since that DOS 6.0 "doublespace/drivespace" nightmare.

 

This forum rawks!
Brad

 

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You're welcome and with Thanks! to Jim for the reminder.

AtomicZombie wrote:
Went from win10 1803.17134.1 to 1803.17134.48
Apr-30'18 to May-8'18

It will have significant issues usually until it reaches SAC or shortly thereafter (approximately 2 or 3 months after a version's first build [RTM])

Issues sometimes arise even in SAC; a Windows 10 operator has ten days after a Windows 10 update to rollback to the previous build.

You're being prudent in holding off Windows Update (you're locking down a functional Windows 10 build)

The older a Windows update is the greater its quality (from a few weeks for an update to be corrected all the way up to Windows 10 Enterprise LTSC)

AtomicZombie wrote:
My invoice for 7 wasted hours over the course of the week has been mailed to Microsoft.
If I was to read your invoice I'd spew some of my coffee onto it and mail it back to you!  wink

WaaS is a WiP.

 


Windows lifecycle fact sheet

Windows 10 release information

MICROSOFT SOFTWARE LICENSE TERMS

WINDOWS OPERATING SYSTEM

Last updated July 2017

https://www.microsoft.com/en-us/Useterms/OEM/Windows/10/Useterms_OEM_Windows_10_English.htm

...

 

12.    Consumer Rights, Regional Variations. 

...

b.      Canada. You may stop receiving updates on your device by turning off Internet access. If and when you re-connect to the Internet, the software will resume checking for and installing updates.

...

 

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AtomicZombie wrote:
The annoying slow build/compile time is back to haunt me after six years!

Six years!  That is one SLLLOOOOWWWWWW compile!

 

We used to have overnight rebuilds on Burroughs B1700 cross-compiler for B800 circa 1975.  But still, that is a factor of 365:1 and computers are supposedly faster now.

 

The sky is falling!  The sky is falling!  Must be that darned Studio.  If this continues, Atmel will be out of business.

 

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Looks like Murphy is fighting back!

 

After the win10 update fixed the slow compile time, I thought all was well.

This morning, I am at my desk trying to use my ICE to load the chip...

 

[ERROR] Dev-48 is not a device context and a bunch of other errors.

 

Nothing moved, nothing changed, and it actually worked ONCE!

 

Now I can't even read the target voltage.

 

So far I have replaced...

 

- Atmega 328p

- Clock

- All wires

- Power supply

- Breadboard

- USB cable

 

No joy.

 

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Problem solved.

 

For the first time in my 20 years of using AVR, I had a chip failure.

In fact, all 3 Atmega328p ICs are toast. Very odd.

I am almost certain that the last 2 I checked were new, as the pins were still too wide for the board.

 

4th one is up and running fine.

Factory defects??

 

I have overclocked others beyond sanity, accidentailly connectd 12v for hours, reversed power, etc.

And yet this is the first IC failure I have seen. Odd as well that I did nothing wrong this time!

 

Anyhow, I am back to coding, and all is well.

 

Thanks for the help.

Brad

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AtomicZombie wrote:
Factory defects??
No

New factory?  Yes (wafer fab)

AtomicZombie wrote:
Odd as well that I did nothing wrong this time!
"It's not me, it's my tools!" (a reminder about you and wall warts wink

Take care and may you have joy!

 


Product Change Notification - GBNG-06LXXH156 - 25 Jun 2018 - CCB 2856 Final Notice: Qualification of Microchip Fabrication site (FAB 4) for selected Atmel products manufactured with the 59.91K process technology.

via https://www.microchip.com/mymicrochip/Reports.aspx?type=cpn&filter=ATmega328P

though am assuming yours are in PDIP due to

... as the pins were still too wide for the board.

There's been a significant amount of PDIP change activity.

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Take the blame???

No way... I need to make myself feel better now that I have only a single AVR left to work with today!

Perhaps later I will accept the truth.

 

Brad

 

I Like to Build Stuff : http://www.AtomicZombie.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AtomicZombie wrote:
Perhaps later I will accept the truth.

Can you "handle the truth"?

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

Pages