-
If you're having problems
loading your CPU into the hardware simulator, try using the built-in memory. It
is written in Java and is much more space-efficient than a hdl file. You may
use whatever sized memory is built-in, you don't have to match the requirement
from project 2 for memory size.
-
The example PC is on
the nova server. See Nova/users/barrj/cs344jb/materials.
-
Presentations will
be at 1:30, Monday, 2 May in Williams 309
(the room the class has met in during the semester).
- I've fixed the figures
in chapter 5, part 2. The updated slides are on the
lecture slides page.
- Exam 2 will take place
on Thursday, 21 April, 6PM in Williams 225.
- There's an interesting
article
here that talks about AltiVect, a vector processing co-processor that's
available on powerPC processors.
- OK, so 8 bits is
big. For problem 1 on PS 7 you may implement a 2-bit adder instead of an
8-bit adder.
- Note that the carry-save
adder requires 4 inputs to work correctly. I've changed problem 3 on problem
set 7 to add 4 numbers.
- Jared and Ali Erkan figured
out a solution to the memory size problem that works for most implementations
of Java:
Open the batch file "HardwareSimulator.bat" in a text editor (like notepad)
and add an argument that indicates the amount of memory that you want XP
to allocate to the Java virtual machine. The original
batch file contains
"@echo off
java -classpath..."
if you change it to
"@echo off
java -Xmx512m -classpath"
XP will allocate 512MB of memory to the Java virtual machine which will be
enough to load your 1K memory file. Change the 512 to whatever the max memory
of your read machine (not the size of the memory that you are creating in HDL).
Given this solution, I'd like you to make your memory 1K now. You have an
extension until Wednesday to turn it in.
- Exam 2 will take place
on Tuesday, 5 April, at 6-8PM in a room TBA.
- I'll accept the
memory part of the project on Monday, 21 March, without penalty.
- There's an interesting
article
here
on how Nils Schneider, a 17-year-old computer science student from Germany,
used some simple hardware hacking to decipher the iPod bootloader and enabled
Linux to run on it.
- Exam 1 will take
place next Tuesday, 1 March, from 6 to 9PM in Williams 222.
- Team assignments
are posted. If you'd like to change teams,
please see me.
- The memory
project has been changed. You only have to build a 1K memory, not a 16K
memory.
- There's an interesting article about RAM at
arstechnica. AnandTech also has a good article on memory
here.
-
The first part of the project, memory, will be due Friday, the
18th of March
(after spring break). See the updated project description.
-
The width problem in the hardware simulator is not about output pins
or input pins. You can have as many (well, I've tested over 35) input and
output pins. The restriction seems to be in the bus size.
A bus seems
to be limited to 16 bits. So you can, for example, have two output busses in
a decoder, each of width 16 for a total output size of 32 bits.
- There's an interesting article on the new IBM "Cell" processor
here at arsTechnica.
- There seems to be
a limit on the number of output pins in a chip in the HardwareSimulator of
16 pins. I'm investigating.
-
I've updated chapter 1 parts 3 and 4. Parts 5-8 have been rearranged, so
you might want to download these slides again.
-
I put a corrected version of the chapter 1 part 2 slides up.