RAM Frequency question

S

Scott Lurndal

joseph said:
Nice. Can you add the FSB information too, so we've got the full
package ?

For the Opteron (only diff for Athlon is a half-size datapath to DRAM
which halves the DRAM bandwidth).

CPU <-> Memory Bandwidth
128-bit data path
200 Mhz clock
Double Data Rate

(200Mhz * 2 * 128 / 8) = 6.4 Gigabytes/sec

Hypertransport Version 2 Bandwidth
16 bit data path
1000 Mhz clock
Double Data Rate
Full Duplex

(1000Mhz * 2 * 16 * 2 / 8) = 8 Gigabytes/sec

PCI Express Bandwidth
x1 512 Megabytes/sec
x4 2 Gigabytes/sec
x8 4 Gigabytes/sec
x16 8 Gigabytes/sec

Gigabit NIC Bandwidth
Bit serial data at 1000 Mhz
10-bit frames

(1000Mhz / 10) = 100 Megabytes/sec

SATA bandwidth
1Gen Link bandwidth 150 Megabytes/sec
2Gen Link bandwidth 300 Megabytes/sec
sustained single drive bandwidth < 100 Megabytes/sec

LPC bandwidth
In the noise.


FibreChannel Bandwidth
Bit serial data at 2 gigabits/sec (SDR) 10-bit frames
Single Data Rate ~200 Megabytes/sec (2Gb)
Double Data Rate ~400 Megabytes/sec (4Gb)
sustained JBOD (single) <100 Megabytes/sec (function of drive RPM and
areal bit density)




Note that the HT can run at 200, 400, 600, 800 or 1000 mhz.

scott
 
Y

Yadayada

Wes Newell said:
Please re-read my post. The 400Mbps IS the data rate of a single line.
That's what I wrote. That's what I meant. And it IS the facts. Actually, I
even wrote "per signal line" to make sure it it wasn't interpreted some
other way. I thought it was very clear.

--

Ok reread, copied, and pasted:

"You've bought into marketing BS. PC3200 ram runs at 200MHz clock, not
400MHz as the marketeers would have you believe. They erroneously use the
data rate of 400Mbps (DDR) as MHz."

That wa the complete post
 
Y

Yadayada

Wes Newell said:
There are no older 400MHz diimms. You've got yourself confused. I never
stated any objection to 400DDR or DDR400. What I object to is them calling
PC3200 ram 400MHz, when it's really 200MHz 400DDR (but not 400Mhz) And
calling PC3200 ram 400Mhz is marketing BS since it's really 200MHz 400DDR.
Again, not 400MHz. Same goes for the FSB of the P4 and The Athlon. It
should be obvious to anyone that's actaully set up many systems when they
set PC3200 ram to 200MHz and not 400MHz. Same for the 400FSB when you set
the FSB to 200MHz and not 400MHz. Anyone that can't tell the diference in
this is just an idiot.

Now all that was true and the real explanantion. It still doesn't change my
original statement, or yours. Sorry you got so bent out of shape Like I
said, I wasn't looking to start a flame war, but your original explanation
was quite obscure and wrong. I guess you just had a bad day that day ;).
 
W

Wes Newell

Ok reread, copied, and pasted:

"You've bought into marketing BS. PC3200 ram runs at 200MHz clock, not
400MHz as the marketeers would have you believe. They erroneously use the
data rate of 400Mbps (DDR) as MHz."

That wa the complete post

And it's all correct. Just what do think is not correct.
 
Y

Yadayada

Wes Newell said:
And it's all correct. Just what do think is not correct.

"They erroneously use the data rate of 400Mbps (DDR) as MHz"

That is not where the name 400DDR comes from. It comes from the "assumed"
speed of the module. Marketing BS? Yes I agree. But it isn't named 400DDR
after it's single line bandwidth. It is called 400DDR because of the fact
that it does 2 words per cycle at 200MHz, making the effective transfer rate
what it would be on 400MHz SDR (if it existed).
 
W

Wes Newell

"They erroneously use the data rate of 400Mbps (DDR) as MHz"
So you actually believe that 400Mbps is the same as 400MHz.
That is not where the name 400DDR comes from. It comes from the
"assumed" speed of the module. Marketing BS? Yes I agree. But it isn't
named 400DDR after it's single line bandwidth. It is called 400DDR
because of the fact that it does 2 words per cycle at 200MHz, making the
effective transfer rate what it would be on 400MHz SDR (if it existed).

Christ, you sure must like to argue, What's the frickin difference where
or how they came up with the number? Personally, I don't know, and I doubt
you really know who came up with it in the first place or how they came
up with it. It erroneous in any case. The Jedec doc someone posted
compared the MHz to the data rate. I guess they are wrong too. In the end,
why it's wrong doesn't really matter. It's just wrong. The one thing you
have done is point it out is that it's wrong figured anyway you want to
figure it.:)
 
Y

Yadayada

Wes Newell said:
So you actually believe that 400Mbps is the same as 400MHz.


Christ, you sure must like to argue, What's the frickin difference where
or how they came up with the number? Personally, I don't know, and I doubt
you really know who came up with it in the first place or how they came
up with it. It erroneous in any case. The Jedec doc someone posted
compared the MHz to the data rate. I guess they are wrong too. In the end,
why it's wrong doesn't really matter. It's just wrong. The one thing you
have done is point it out is that it's wrong figured anyway you want to
figure it.:)

Like I said in my original statement, I didn't want it to turn into a flame
war, but I should have known better with who I was correcting. :) It was
quite obvious you were in a very bad mood when you made the original post
anyway, so I guess I should have left well enough alone. :p
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top