Assembly Isn't That Bad:

[davidthefat]

So I thought assembly would be a very hard thing to learn, but surprisingly, it's not bad as all. It is not scary, or anything. I understand it. So why are people so scared of it? BTW, I have only done basic operations and using inline assembly. May be using it more will prove to be a pain. Also personally, I think the AT&T syntax makes more sense.

[Ginto8]

Assembly isn't hard per se, but if you want to do more complex things, you end up doing a lot more work to do simple things. I haven't done a ton of stuff with assembly, but C at least is probably the lowest amount of abstraction that I'm comfortable with for most things.

[davidthefat]

Oh whether I like it or not is not a choice . I'll eventually have to learn it anyway. I plan on double majoring in Computer Engineering and Computer Science and taking extra theoretical physics courses (don't say it is a stupid idea, I want to work on quantum computers, robotics, artificial intelligence, computer architectures, and just anything computer research related) when I go to college. Personally, I am the type that rather take a lego making machine over a ton of legos. If you know what I mean.

[short]

So why are people so scared of it? BTW, I have only done basic operations and using inline assembly.

This is why:

Code: Select all

;***********************************************************
;*
;*	Lab6.asm
;*
;*	USART RECEIVER
;*
;***********************************************************
;*
;*	 Author: Benjamin Adamson
;*	 		 
;*	   Date: February 26, 2011
;*
;***********************************************************


;************************************************************
;* Include files
;************************************************************

.include "m128def.inc"			; Include definition file

;***********************************************************
;*	Internal Register Definitions and Constants
;***********************************************************
.def	mpr = r16				; Multi-purpose register 

.def	waitcnt = r17			; Wait Loop Counter
.def	ilcnt = r18				; Inner Loop Counter
.def	olcnt = r19				; Outer Loop Counter
.def 	deviceid = r22			; Device ID
.def 	actioncode = r23		; Action Code
.def 	temp = r24				; Temporary register
.def 	flag = r25				; flag register to signal wheter to ignore commands

.equ	WTime = 100				; Time to wait in wait loop
.equ	WTimeReverse = 200		; Time to wait in reverse wait loop

.equ	WskrR = 0				; Right Whisker Input Bit
.equ	WskrL = 1				; Left Whisker Input Bit
.equ	EngEnR = 4				; Right Engine Enable Bit
.equ	EngEnL = 7				; Left Engine Enable Bit
.equ	EngDirR = 5				; Right Engine Direction Bit
.equ	EngDirL = 6				; Left Engine Direction Bit

;/////////////////////////////////////////////////////////////
;These macros are the values to make the TekBot Move.
;/////////////////////////////////////////////////////////////

.equ	MovFwd = (1<<EngDirR|1<<EngDirL); Move Forwards Command
.equ	MovBck = $00					; Move Backwards Command
.equ	TurnR = (1<<EngDirL)			; Turn Right Command
.equ	TurnL = (1<<EngDirR)			; Turn Left Command
.equ	Halt = (1<<EngEnR|1<<EngEnL)	; Halt Command


;***********************************************************
;*	Start of Code Segment
;***********************************************************
.cseg							; Beginning of code segment

;-----------------------------------------------------------
; Interrupt Vectors
;-----------------------------------------------------------
.org	$0000					; Beginning of IVs
		rjmp 	INIT			; Reset interrupt

; Set up the interrupt vectors for the interrupts
.org	$0002					; IRQ0 handler
		rcall	HitRight		; Call hit right function
		reti					; Return from interrupt

.org 	$0004					; IRQ1 Handler
		rcall HitLeft			; Call hit left function
		reti					; Return from interrupt

.org 	$003C					; USART1, Rx Complete Handler
		rcall DecodeAndExecute	; Call DecodeAndExecute function
		reti					; Return from interrupt 

.org	$0046					; End of Interrupt Vectors

;--------------------------------------------------------------
; Program Initialization
;--------------------------------------------------------------
INIT:
		; Initilize the Stack Pointer
		ldi		mpr, low(RAMEND)	; Place low RAMEND into mpr
		out		SPL, mpr			; Load SPL with low byte of RAMEND
		ldi		mpr, high(RAMEND)	; place high RAMEND into mpr
		out		SPH, mpr			; Load SPH with high byte of RAMEND

				; Setup port B pins 11110000 for output
		ldi		mpr, (1<<EngEnL)|(1<<EngEnR)|(1<<EngDirR)|(1<<EngDirL)
		out		DDRB, mpr		; Set the DDRB register for output

		; Initialize Port B for output
		ldi		mpr, $00		; Initialize Port B for outputs
		out		PORTB, mpr		; Port B outputs low

		; Initialize Port D for inputs
		ldi		mpr, (0<<WskrL)|(0<<WskrR)		; Set Port D Directional Register
		out		DDRD, mpr						; for inputs (0 => input, 1 => output)


		ldi		mpr, (1<<WskrL)|(1<<WskrR)		; Initialize Port D for inputs
		out		PORTD, mpr						; with Tri-State



		; Initialize external interrupts!
		; Set the Interrupt Sense Control to falling level edge
		ldi mpr, (1<<ISC01)|(0<<ISC00)|(1<<ISC11)|(0<<ISC10)
		sts EICRA, mpr	; Set INT0 & INT1 to trigger on falling edge!

		ldi mpr, $00	; Load all zeros into mpr
		sts EICRB, mpr 	; Use sts, EICRB in extended I/O space!

		; External Interrupt Mask!
		ldi mpr, (1<<INT0)|(1<<INT1); Load 00000011 into mpr
		out EIMSK, mpr				; Load mpr into EIMSK

USART_Init:
		; Set frame format: 8data, 2stop bit
		ldi mpr, (1<<USBS1)|(3<<UCSZ10)
		sts UCSR1C,mpr
		; NOTE if anything weird happens verify this initialize USART code.

		; Set baud rate
		ldi mpr, 0b00000001			; high byte of '2400'
		sts UBRR1H, mpr

		ldi mpr, 0b10100000			; low byte of '2400'
		sts UBRR1L, mpr

		; Enable receiver and transmitter
		ldi mpr, (1<<RXEN1)|(1<<TXEN1)
		sts UCSR1B, mpr


SETFLAG:
		clr flag

SETUP_MOVEMENT:
		ldi actioncode, MovFwd

		sei							; Enable global interrupts


;-----------------------------------------------------------
; Main Program
;-----------------------------------------------------------
MAIN:	; The Main program

		; Initialize TekBot Foward Movement
		out		PORTB, actioncode		; Send command to motors

		rjmp	MAIN			; Create an infinite while loop to signify the 
								; end of the program.

;****************************************************************
;* Subroutines and Functions
;****************************************************************

;----------------------------------------------------------------
; Sub:	HitRight
; Desc:	Handles functionality of the TekBot when the right whisker
;		is triggered.
;----------------------------------------------------------------
HitRight:
		cli					; Disable global interrupts!
		push	mpr			; Save mpr register
		push	waitcnt		; Save wait register
		in		mpr, SREG	; Save program state
		push	mpr			; Save mpr state

		; Move Backwards for a second
		ldi		mpr, MovBck				; Load Move Backwards command
		out		PORTB, mpr				; Send command to port
		ldi		waitcnt, WTimeReverse	; Wait for 2 seconds
		;rcall	Wait					; Call wait function

		; Turn left for a second
		ldi		mpr, TurnL		; Load Turn Left Command
		out		PORTB, mpr		; Send command to port
		ldi		waitcnt, WTime	; Wait for 1 second
		;rcall	Wait			; Call wait function

		; Move Forward again	
		ldi		mpr, MovFwd	; Load Move Forwards command
		out		PORTB, mpr	; Send command to port

		pop		mpr			; Restore program state
		out		SREG, mpr	; SREG was stored into mpr, so use mpr to restore SREG
		pop		waitcnt		; Restore wait register
		pop		mpr			; Restore mpr
		sei					; Enable global interrupts!
		ret					; Return from subroutine

;----------------------------------------------------------------
; Sub:	HitLeft
; Desc:	Handles functionality of the TekBot when the left whisker
;		is triggered.
;----------------------------------------------------------------
HitLeft:
		cli					; Disable global interrupts
		push	mpr			; Save mpr register
		push	waitcnt		; Save wait register
		in		mpr, SREG	; Save program state
		push	mpr			; Save mpr state

		; Move Backwards for a second
		ldi		mpr, MovBck				; Load Move Backwards command
		out		PORTB, mpr				; Send command to port
		ldi		waitcnt, WTimeReverse	; Wait for 2 seconds
		;rcall	Wait					; Call wait function

		; Turn right for a second
		ldi		mpr, TurnR		; Load Turn Left Command
		out		PORTB, mpr		; Send command to port
		ldi		waitcnt, WTime	; Wait for 1 second
		;rcall	Wait			; Call wait function

		; Move Forward again	
		ldi		mpr, MovFwd	; Load Move Forwards command
		out		PORTB, mpr	; Send command to port

		pop		mpr			; Restore program state
		out		SREG, mpr	; SREG was stored into mpr, so use mpr to restore SREG
		pop		waitcnt		; Restore wait register
		pop		mpr			; Restore mpr
		sei					; Enable global interrupts!
		ret					; Return from subroutine

;----------------------------------------------------------------
; Sub:	Wait
; Desc:	A wait loop that is 16 + 159975*waitcnt cycles or roughly 
;		waitcnt*10ms.  Just initialize wait for the specific amount 
;		of time in 10ms intervals. Here is the general eqaution
;		for the number of clock cycles in the wait loop:
;			((3 * ilcnt + 3) * olcnt + 3) * waitcnt + 13 + call
;----------------------------------------------------------------
Wait:
		push	waitcnt			; Save wait register
		push	ilcnt			; Save ilcnt register
		push	olcnt			; Save olcnt register

Loop:	ldi		olcnt, 224		; load olcnt register
OLoop:	ldi		ilcnt, 237		; load ilcnt register
ILoop:	dec		ilcnt			; decrement ilcnt
		brne	ILoop			; Continue Inner Loop
		dec		olcnt		; decrement olcnt
		brne	OLoop			; Continue Outer Loop
		dec		waitcnt		; Decrement wait 
		brne	Loop			; Continue Wait loop	

		pop		olcnt		; Restore olcnt register
		pop		ilcnt		; Restore ilcnt register
		pop		waitcnt		; Restore wait register
		ret				; Return from subroutine

;----------------------------------------------------------------
; Sub:	DecodeAndExecute
; Desc:	Handles functionality of the TekBot when a transmission is received.
;----------------------------------------------------------------
DecodeAndExecute:
		;cli					; clear interrupts
		push mpr			; save state mpr
		;push r20			; save r20 state
		;push r21			; save r21 state
		
		; Get status and 9th bit, then data from buffer
		;lds r20, UCSR1A		; status?
		;lds r21, UCSR1B		; 9th bit?
		lds mpr, UDR1		; data?
		out PORTB, mpr
		pop mpr
		ret


		; If error, return -1
		;andi r21,(1<<FE1)|(1<<DOR1)|(1<<UPE1)
		;breq USART_ReceiveNoError
		;ldi r20, HIGH(-1)
		;ldi mpr, LOW(-1)

USART_ReceiveNoError:

		; Filter the 9th bit, then return
		;lsr r20
		;andi r20, 0x01

		; if the device id has not been read
		;lds temp, (UCSR1A << RXC1) 		; setup for buffer flushing

		cpi flag, 0x00				
		brne ACTION


DEVICE_ID_RECV:
		; check to see if the device ID is equal to this receivers ID
		; if it is, store
		lds deviceid, UDR1
		ldi flag, 0x01
		rjmp FLUSH_RECEIVE_BUFFER

ACTION:
		lds actioncode, (UDR1<<1)
		ldi flag, 0x00

FLUSH_RECEIVE_BUFFER:

	;	sbis UCSR1A, RXC1

		andi temp, 0b10000000
		brne END

		lds r16, UDR1
		rjmp FLUSH_RECEIVE_BUFFER
END:
		pop r21				; restore r21
		pop r20				; restore r20
		pop mpr				; restore mpr
		sei 				; reenable interrupts
		ret

It's programs like this that make me hate assembly. I hate that my university has forced me to learn three assembly languages thus far into my bachelors. The code above is for an AVR atmega128 chip to power the infrared receiver, using hardware interrupts. It's difficult as hell to remember what everything is on a piece of hardware I have no interest in ever programming for again.

edit: I'm not sure why copy/paste screwed with the tab indenting.

[davidthefat]

So why are people so scared of it? BTW, I have only done basic operations and using inline assembly.

This is why:

Code: Select all

;***********************************************************
;*
;*	Lab6.asm
;*
;*	USART RECEIVER
;*
;***********************************************************
;*
;*	 Author: Benjamin Adamson
;*	 		 
;*	   Date: February 26, 2011
;*
;***********************************************************


;************************************************************
;* Include files
;************************************************************

.include "m128def.inc"			; Include definition file

;***********************************************************
;*	Internal Register Definitions and Constants
;***********************************************************
.def	mpr = r16				; Multi-purpose register 

.def	waitcnt = r17			; Wait Loop Counter
.def	ilcnt = r18				; Inner Loop Counter
.def	olcnt = r19				; Outer Loop Counter
.def 	deviceid = r22			; Device ID
.def 	actioncode = r23		; Action Code
.def 	temp = r24				; Temporary register
.def 	flag = r25				; flag register to signal wheter to ignore commands

.equ	WTime = 100				; Time to wait in wait loop
.equ	WTimeReverse = 200		; Time to wait in reverse wait loop

.equ	WskrR = 0				; Right Whisker Input Bit
.equ	WskrL = 1				; Left Whisker Input Bit
.equ	EngEnR = 4				; Right Engine Enable Bit
.equ	EngEnL = 7				; Left Engine Enable Bit
.equ	EngDirR = 5				; Right Engine Direction Bit
.equ	EngDirL = 6				; Left Engine Direction Bit

;/////////////////////////////////////////////////////////////
;These macros are the values to make the TekBot Move.
;/////////////////////////////////////////////////////////////

.equ	MovFwd = (1<<EngDirR|1<<EngDirL); Move Forwards Command
.equ	MovBck = $00					; Move Backwards Command
.equ	TurnR = (1<<EngDirL)			; Turn Right Command
.equ	TurnL = (1<<EngDirR)			; Turn Left Command
.equ	Halt = (1<<EngEnR|1<<EngEnL)	; Halt Command


;***********************************************************
;*	Start of Code Segment
;***********************************************************
.cseg							; Beginning of code segment

;-----------------------------------------------------------
; Interrupt Vectors
;-----------------------------------------------------------
.org	$0000					; Beginning of IVs
		rjmp 	INIT			; Reset interrupt

; Set up the interrupt vectors for the interrupts
.org	$0002					; IRQ0 handler
		rcall	HitRight		; Call hit right function
		reti					; Return from interrupt

.org 	$0004					; IRQ1 Handler
		rcall HitLeft			; Call hit left function
		reti					; Return from interrupt

.org 	$003C					; USART1, Rx Complete Handler
		rcall DecodeAndExecute	; Call DecodeAndExecute function
		reti					; Return from interrupt 

.org	$0046					; End of Interrupt Vectors

;--------------------------------------------------------------
; Program Initialization
;--------------------------------------------------------------
INIT:
		; Initilize the Stack Pointer
		ldi		mpr, low(RAMEND)	; Place low RAMEND into mpr
		out		SPL, mpr			; Load SPL with low byte of RAMEND
		ldi		mpr, high(RAMEND)	; place high RAMEND into mpr
		out		SPH, mpr			; Load SPH with high byte of RAMEND

				; Setup port B pins 11110000 for output
		ldi		mpr, (1<<EngEnL)|(1<<EngEnR)|(1<<EngDirR)|(1<<EngDirL)
		out		DDRB, mpr		; Set the DDRB register for output

		; Initialize Port B for output
		ldi		mpr, $00		; Initialize Port B for outputs
		out		PORTB, mpr		; Port B outputs low

		; Initialize Port D for inputs
		ldi		mpr, (0<<WskrL)|(0<<WskrR)		; Set Port D Directional Register
		out		DDRD, mpr						; for inputs (0 => input, 1 => output)


		ldi		mpr, (1<<WskrL)|(1<<WskrR)		; Initialize Port D for inputs
		out		PORTD, mpr						; with Tri-State



		; Initialize external interrupts!
		; Set the Interrupt Sense Control to falling level edge
		ldi mpr, (1<<ISC01)|(0<<ISC00)|(1<<ISC11)|(0<<ISC10)
		sts EICRA, mpr	; Set INT0 & INT1 to trigger on falling edge!

		ldi mpr, $00	; Load all zeros into mpr
		sts EICRB, mpr 	; Use sts, EICRB in extended I/O space!

		; External Interrupt Mask!
		ldi mpr, (1<<INT0)|(1<<INT1); Load 00000011 into mpr
		out EIMSK, mpr				; Load mpr into EIMSK

USART_Init:
		; Set frame format: 8data, 2stop bit
		ldi mpr, (1<<USBS1)|(3<<UCSZ10)
		sts UCSR1C,mpr
		; NOTE if anything weird happens verify this initialize USART code.

		; Set baud rate
		ldi mpr, 0b00000001			; high byte of '2400'
		sts UBRR1H, mpr

		ldi mpr, 0b10100000			; low byte of '2400'
		sts UBRR1L, mpr

		; Enable receiver and transmitter
		ldi mpr, (1<<RXEN1)|(1<<TXEN1)
		sts UCSR1B, mpr


SETFLAG:
		clr flag

SETUP_MOVEMENT:
		ldi actioncode, MovFwd

		sei							; Enable global interrupts


;-----------------------------------------------------------
; Main Program
;-----------------------------------------------------------
MAIN:	; The Main program

		; Initialize TekBot Foward Movement
		out		PORTB, actioncode		; Send command to motors

		rjmp	MAIN			; Create an infinite while loop to signify the 
								; end of the program.

;****************************************************************
;* Subroutines and Functions
;****************************************************************

;----------------------------------------------------------------
; Sub:	HitRight
; Desc:	Handles functionality of the TekBot when the right whisker
;		is triggered.
;----------------------------------------------------------------
HitRight:
		cli					; Disable global interrupts!
		push	mpr			; Save mpr register
		push	waitcnt		; Save wait register
		in		mpr, SREG	; Save program state
		push	mpr			; Save mpr state

		; Move Backwards for a second
		ldi		mpr, MovBck				; Load Move Backwards command
		out		PORTB, mpr				; Send command to port
		ldi		waitcnt, WTimeReverse	; Wait for 2 seconds
		;rcall	Wait					; Call wait function

		; Turn left for a second
		ldi		mpr, TurnL		; Load Turn Left Command
		out		PORTB, mpr		; Send command to port
		ldi		waitcnt, WTime	; Wait for 1 second
		;rcall	Wait			; Call wait function

		; Move Forward again	
		ldi		mpr, MovFwd	; Load Move Forwards command
		out		PORTB, mpr	; Send command to port

		pop		mpr			; Restore program state
		out		SREG, mpr	; SREG was stored into mpr, so use mpr to restore SREG
		pop		waitcnt		; Restore wait register
		pop		mpr			; Restore mpr
		sei					; Enable global interrupts!
		ret					; Return from subroutine

;----------------------------------------------------------------
; Sub:	HitLeft
; Desc:	Handles functionality of the TekBot when the left whisker
;		is triggered.
;----------------------------------------------------------------
HitLeft:
		cli					; Disable global interrupts
		push	mpr			; Save mpr register
		push	waitcnt		; Save wait register
		in		mpr, SREG	; Save program state
		push	mpr			; Save mpr state

		; Move Backwards for a second
		ldi		mpr, MovBck				; Load Move Backwards command
		out		PORTB, mpr				; Send command to port
		ldi		waitcnt, WTimeReverse	; Wait for 2 seconds
		;rcall	Wait					; Call wait function

		; Turn right for a second
		ldi		mpr, TurnR		; Load Turn Left Command
		out		PORTB, mpr		; Send command to port
		ldi		waitcnt, WTime	; Wait for 1 second
		;rcall	Wait			; Call wait function

		; Move Forward again	
		ldi		mpr, MovFwd	; Load Move Forwards command
		out		PORTB, mpr	; Send command to port

		pop		mpr			; Restore program state
		out		SREG, mpr	; SREG was stored into mpr, so use mpr to restore SREG
		pop		waitcnt		; Restore wait register
		pop		mpr			; Restore mpr
		sei					; Enable global interrupts!
		ret					; Return from subroutine

;----------------------------------------------------------------
; Sub:	Wait
; Desc:	A wait loop that is 16 + 159975*waitcnt cycles or roughly 
;		waitcnt*10ms.  Just initialize wait for the specific amount 
;		of time in 10ms intervals. Here is the general eqaution
;		for the number of clock cycles in the wait loop:
;			((3 * ilcnt + 3) * olcnt + 3) * waitcnt + 13 + call
;----------------------------------------------------------------
Wait:
		push	waitcnt			; Save wait register
		push	ilcnt			; Save ilcnt register
		push	olcnt			; Save olcnt register

Loop:	ldi		olcnt, 224		; load olcnt register
OLoop:	ldi		ilcnt, 237		; load ilcnt register
ILoop:	dec		ilcnt			; decrement ilcnt
		brne	ILoop			; Continue Inner Loop
		dec		olcnt		; decrement olcnt
		brne	OLoop			; Continue Outer Loop
		dec		waitcnt		; Decrement wait 
		brne	Loop			; Continue Wait loop	

		pop		olcnt		; Restore olcnt register
		pop		ilcnt		; Restore ilcnt register
		pop		waitcnt		; Restore wait register
		ret				; Return from subroutine

;----------------------------------------------------------------
; Sub:	DecodeAndExecute
; Desc:	Handles functionality of the TekBot when a transmission is received.
;----------------------------------------------------------------
DecodeAndExecute:
		;cli					; clear interrupts
		push mpr			; save state mpr
		;push r20			; save r20 state
		;push r21			; save r21 state
		
		; Get status and 9th bit, then data from buffer
		;lds r20, UCSR1A		; status?
		;lds r21, UCSR1B		; 9th bit?
		lds mpr, UDR1		; data?
		out PORTB, mpr
		pop mpr
		ret


		; If error, return -1
		;andi r21,(1<<FE1)|(1<<DOR1)|(1<<UPE1)
		;breq USART_ReceiveNoError
		;ldi r20, HIGH(-1)
		;ldi mpr, LOW(-1)

USART_ReceiveNoError:

		; Filter the 9th bit, then return
		;lsr r20
		;andi r20, 0x01

		; if the device id has not been read
		;lds temp, (UCSR1A << RXC1) 		; setup for buffer flushing

		cpi flag, 0x00				
		brne ACTION


DEVICE_ID_RECV:
		; check to see if the device ID is equal to this receivers ID
		; if it is, store
		lds deviceid, UDR1
		ldi flag, 0x01
		rjmp FLUSH_RECEIVE_BUFFER

ACTION:
		lds actioncode, (UDR1<<1)
		ldi flag, 0x00

FLUSH_RECEIVE_BUFFER:

	;	sbis UCSR1A, RXC1

		andi temp, 0b10000000
		brne END

		lds r16, UDR1
		rjmp FLUSH_RECEIVE_BUFFER
END:
		pop r21				; restore r21
		pop r20				; restore r20
		pop mpr				; restore mpr
		sei 				; reenable interrupts
		ret

It's programs like this that make me hate assembly. I hate that my university has forced me to learn three assembly languages thus far into my bachelors. The code above is for an AVR atmega128 chip to power the infrared receiver, using hardware interrupts. It's difficult as hell to remember what everything is on a piece of hardware I have no interest in ever programming for again.

edit: I'm not sure why copy/paste screwed with the tab indenting.

Eh, I guess I'll have to learn to literally comment every line if I don't want nightmares every night.

[short]

Comments on every line are a requirement of my prof something about making sure we understand what we wrote

[MrDeathNote]

I plan on double majoring in Computer Engineering and Computer Science and taking extra theoretical physics courses (don't say it is a stupid idea, I want to work on quantum computers, robotics, artificial intelligence, computer architectures, and just anything computer research related) when I go to college.

I would never say thats a stupid idea. I was a theoretical physics student before I was a Comp sci student. I can tell you though that you aren't likely to do much work with quantum computing. All the courses I did were based around quantum mechanics, partical physics, thermodynamics, relativity and a lot more that I can't even remember now. Also you end up doing a lot of pure maths courses.

[N64vSNES]

If C is a back rub, assembly is a knife in the back.

[Falco Girgis]

There are only two reasons that assembly can ever be hard:

1) It is a nightmare to do anything substantial in assembly
2) It requires intricate knowledge of the CPU and hardware architecture. So naturally the more complex the architecture, the more nightmarish the instruction set and tasks required to achieve a goal.

I'm guessing you were introduced to moving things around in registers and calling simple math instructions. Trust me, there's a shitload more to it than that.

Double majoring in Computer Science and computer engineering? Sorry, but I honestly would say that's a stupid idea. The two already overlap in almost half of the spectrum. I have never heard of a computer engineer not being able to get a job as a computer scientist OR an electrical engineer. CPE is broader than CS. Yes, you go into software less, but you learn a lot more about architecture and hardware.

Since a CPE can get a job as a CS, what is the point? You're wasting precious class money and hours. If you really seek a sincere pursuit of knowledge, this is still not a very intelligent alternative. I will ALWAYS recommend getting a master's as opposed to a double major. Masters are respected, result in pay increases, and are utilized in the work place. A double major is almost useless.

If you're regretting not being CPE (and choosing CS), just finish your degree and get a masters in CPE. It's really not that much of a change. And guess what? You haven't wasted another two years of your time and money for another undergraduate degree that ... sorry to say... is useless...

[davidthefat]

There are only two reasons that assembly can ever be hard:

1) It is a nightmare to do anything substantial in assembly
2) It requires intricate knowledge of the CPU and hardware architecture. So naturally the more complex the architecture, the more nightmarish the instruction set and tasks required to achieve a goal.

I'm guessing you were introduced to moving things around in registers and calling simple math instructions. Trust me, there's a shitload more to it than that.

Double majoring in Computer Science and computer engineering? Sorry, but I honestly would say that's a stupid idea. The two already overlap in almost half of the spectrum. I have never heard of a computer engineer not being able to get a job as a computer scientist OR an electrical engineer. CPE is broader than CS. Yes, you go into software less, but you learn a lot more about architecture and hardware.

Since a CPE can get a job as a CS, what is the point? You're wasting precious class money and hours. If you really seek a sincere pursuit of knowledge, this is still not a very intelligent alternative. I will ALWAYS recommend getting a master's as opposed to a double major. Masters are respected, result in pay increases, and are utilized in the work place. A double major is almost useless.

If you're regretting not being CPE (and choosing CS), just finish your degree and get a masters in CPE. It's really not that much of a change. And guess what? You haven't wasted another two years of your time and money for another undergraduate degree that ... sorry to say... is useless...

I'm aiming for the doctorate just sayin'. I want my job to be very research intensive, like at NASA, IBM, Intel, hell, even AMD. Or some other place like Boeing, Lockheed Martin and ect.

[TheBuzzSaw]

Well yeah, at its most fundamental level, assembly is really quite easy. There are not that many instructions. It is probably easier to learn the entirety of a particular form of assembly than it is to learn all of C/C++. As people stated in the thread, it's when you start trying to do higher level logic that it breaks down and quickly becomes "hard". A good example of this would be trying to implement C++ objects in C. It's not that hard to implement constructors, destructors, virtual functions, etc. but good bloody luck remembering to always call the constructor at the beginning, the destructor at the end, etc.

I learned assembly for my computer architecture course. It was a wonderful exercise, but I would never ever use assembly on anything substantial today.

[Falco Girgis]

There are only two reasons that assembly can ever be hard:

1) It is a nightmare to do anything substantial in assembly
2) It requires intricate knowledge of the CPU and hardware architecture. So naturally the more complex the architecture, the more nightmarish the instruction set and tasks required to achieve a goal.

I'm guessing you were introduced to moving things around in registers and calling simple math instructions. Trust me, there's a shitload more to it than that.

Double majoring in Computer Science and computer engineering? Sorry, but I honestly would say that's a stupid idea. The two already overlap in almost half of the spectrum. I have never heard of a computer engineer not being able to get a job as a computer scientist OR an electrical engineer. CPE is broader than CS. Yes, you go into software less, but you learn a lot more about architecture and hardware.

Since a CPE can get a job as a CS, what is the point? You're wasting precious class money and hours. If you really seek a sincere pursuit of knowledge, this is still not a very intelligent alternative. I will ALWAYS recommend getting a master's as opposed to a double major. Masters are respected, result in pay increases, and are utilized in the work place. A double major is almost useless.

If you're regretting not being CPE (and choosing CS), just finish your degree and get a masters in CPE. It's really not that much of a change. And guess what? You haven't wasted another two years of your time and money for another undergraduate degree that ... sorry to say... is useless...

I'm aiming for the doctorate just sayin'. I want my job to be very research intensive, like at NASA, IBM, Intel, hell, even AMD. Or some other place like Boeing, Lockheed Martin and ect.

Great! Seriously! I would like a phD as well... So why would you waste time and money getting an EXTREMELY redundant double major like computer engineering and computer science when you could be furthering your education and obtaining a higher, more respected, more useful degree? There was a time when I wanted to double major in either mathematics or electrical engineering with my computer engineering degree. Just because I "loved the subject and wanted to pursue knowledge." Once you get into college and are introduced to the work-force, you realize that this is really silly. A dual major is literally a waste of time and money. You do NOT need a piece of paper saying that you know something to know something.

Also, I have another bit of bad news for you. 90% of jobs are NOT research intensive, unless you work in R&D which is VERY rare. Only the best of the best work there, and you have to work you way up. How do I know what I'm talking about? I worked as a government contractor for Stanley Associates (fortune 100 engineering company) for the Army on the Redstone arsenal for about 2 years. I literally just interviewed with Boeing and Lockheed Martin (I graduate in two months), my father works as an engineer for NASA, and I have friends who work for Boeing, Lockheed, and all of those other big, private sector companies. It's just a sad fact that "research"--especially for newly graduates with nothing more than an undergraduate degree--is not marketable, not profitable, and not in a company's own selfish self interest.

If you are interested in an earnest pursuit of knowledge, aim for the same goal I'm trying to reach. Get hired by one of those high-paying, high-profile jobs. Do useful, meaningful work for them (with one undergrad), and 90% of them pay for your entire Masters and phD for free. If you "genuinely" want to get into research, your future almost certainly must be within academia. And trust me, it's at the GRADUATE level pursuing a Masters or phD, not as an UNDERGRADUATE wasting more time on redundant bachelors. The difference between undergraduate and graduate is literally the pursuit of a piece of paper to get you a job versus the pursuit of knowledge as an end to itself. You have much to learn, grasshopper.

[ParticleGames]

Dual majors aren't bad, but if your majors are as redundant as his then it's just silly. (like you said, Gyrovorbis)

[davidthefat]

There are only two reasons that assembly can ever be hard:

1) It is a nightmare to do anything substantial in assembly
2) It requires intricate knowledge of the CPU and hardware architecture. So naturally the more complex the architecture, the more nightmarish the instruction set and tasks required to achieve a goal.

I'm guessing you were introduced to moving things around in registers and calling simple math instructions. Trust me, there's a shitload more to it than that.

Double majoring in Computer Science and computer engineering? Sorry, but I honestly would say that's a stupid idea. The two already overlap in almost half of the spectrum. I have never heard of a computer engineer not being able to get a job as a computer scientist OR an electrical engineer. CPE is broader than CS. Yes, you go into software less, but you learn a lot more about architecture and hardware.

Since a CPE can get a job as a CS, what is the point? You're wasting precious class money and hours. If you really seek a sincere pursuit of knowledge, this is still not a very intelligent alternative. I will ALWAYS recommend getting a master's as opposed to a double major. Masters are respected, result in pay increases, and are utilized in the work place. A double major is almost useless.

If you're regretting not being CPE (and choosing CS), just finish your degree and get a masters in CPE. It's really not that much of a change. And guess what? You haven't wasted another two years of your time and money for another undergraduate degree that ... sorry to say... is useless...

I'm aiming for the doctorate just sayin'. I want my job to be very research intensive, like at NASA, IBM, Intel, hell, even AMD. Or some other place like Boeing, Lockheed Martin and ect.

Great! Seriously! I would like a phD as well... So why would you waste time and money getting an EXTREMELY redundant double major like computer engineering and computer science when you could be furthering your education and obtaining a higher, more respected, more useful degree? There was a time when I wanted to double major in either mathematics or electrical engineering with my computer engineering degree. Just because I "loved the subject and wanted to pursue knowledge." Once you get into college and are introduced to the work-force, you realize that this is really silly. A dual major is literally a waste of time and money. You do NOT need a piece of paper saying that you know something to know something.

Also, I have another bit of bad news for you. 90% of jobs are NOT research intensive, unless you work in R&D which is VERY rare. Only the best of the best work there, and you have to work you way up. How do I know what I'm talking about? I worked as a government contractor for Stanley Associates (fortune 100 engineering company) for the Army on the Redstone arsenal for about 2 years. I literally just interviewed with Boeing and Lockheed Martin (I graduate in two months), my father works as an engineer for NASA, and I have friends who work for Boeing, Lockheed, and all of those other big, private sector companies. It's just a sad fact that "research"--especially for newly graduates with nothing more than an undergraduate degree--is not marketable, not profitable, and not in a company's own selfish self interest.

If you are interested in an earnest pursuit of knowledge, aim for the same goal I'm trying to reach. Get hired by one of those high-paying, high-profile jobs. Do useful, meaningful work for them (with one undergrad), and 90% of them pay for your entire Masters and phD for free. If you "genuinely" want to get into research, your future almost certainly must be within academia. And trust me, it's at the GRADUATE level pursuing a Masters or phD, not as an UNDERGRADUATE wasting more time on redundant bachelors. The difference between undergraduate and graduate is literally the pursuit of a piece of paper to get you a job versus the pursuit of knowledge as an end to itself. You have much to learn, grasshopper.

Thanks for the info. Yes, I do have much to learn.

[MrDeathNote]

Double majoring in Computer Science and computer engineering? Sorry, but I honestly would say that's a stupid idea

Just to clarify I meant the theoretical physics part wasn't stupid not the double major, I agree that those two subjects are far too interrelated and doing both is redundant.