ANN: Amontec JTAGkey programs XC4VLX25 at 2.8s

ON NEXT MONDAY : 17-JUNE-2006

Amontec will provide the ?how-to? program via a XILINX VIRTEX XC4VLX25

7.9Mbits bit stream) at 2.8 seconds using the Amontec JTAGkey !

On next Monday, your Amontec JTAG key will be close to the speed of a Xilinx Platform Cable USB for programming any FPGA and CPLD vendors (Altera Xilinx Lattice Cypress ...)

Come back next Monday on

formatting link
!

Laurent

Reply to
Amontec, Larry
Loading thread data ...

Hi Laurent

when making public announcements its generically a good idea not to lie.I am quoting your website:

"On next Monday, your Amontec JTAGkey will receive the speed of a Xilinx Platform Cable USB for programming any FPGA (Altera Xilinx Lattice Cypress ...)"

Platform USB Cable is USB HS Device, with dedicated CPLD for JTAG. It can support 24Mbit JTAG clock Amontec Key, is FT2232 a USB FS device with kinda support for JTAG with max bit rate of 6MBit

the fact that it may take same time to program some FPGA under some condition with those 2 programmers doesnt really mean that amonteckey is as same speed as xilinx platform usb.

the very next update to xilinx cable FW and or CPLD may increase the xilinx speed to the theoreatical maximum, that is at 4 times the speed any FT2232 MPSSE solution (without extra CPLD) can ever support.

4 times is not "same" to me.

or have you find a way to convert FT2232 to HS device? or are you offering a HS USB solution to all customers?

sorry, but I preffer when technical matters are explained correctly.

Antti

Reply to
Antti

That depends on if it's true in a specific case, or if it's true in general.

The theoretical speed of well designed hardware is unimportant, what matters is the speed of the available-to-users hardware + software system. Unless of course you end up limited by signal integrity issues instead...

Reply to
cs_posting

Hi Laurent,

I assumed you meant next monday, as "monday next week" from current day, this would then be TODAY but on your website it reads: "ON NEXT MONDAY : 17-JUNE-2006" this is monday a year ago??

so where is this 2.8 second HOWTO "coming next monday" is it going to be online today ? or was it online year ago? if it was year ago then its really old news.

hm.. ok, the monday is almos over, in japan its way past working hours already, and the HOWTO is not yet available, hm should we all look your website in june

2006 ?? ah, maybe I should use webarchive.org to fetch the old pages?

Antti is still hoping to see how amontec FS device works faster then HS USB from Xilinx

Reply to
Antti

Amontec is claiming "7.9Mbits bit stream) at 2.8 seconds" which works out to 2.82 Mbit/second, well less than the data transfer capability of usb full speed (some 12 mbit/s if I recall correctly).

Do you have evidence that xilinx is actually running faster, instead of merely using a usb chip that is theoretically capable of going much faster than it's actually being used in their product?

Reply to
cs_posting

so far nobody can actually compare the 2 cables under same condictions because Amontec promised their "2.8 second howto" to be available "next monday" what would be yesterday, and as of the moment of writing Amontec website still has no actual updated info about this claimed "speed improvment" :(

if/when Amontec fullfills their promise then it will be possible to compare the speeds

Antti

Reply to
Antti

Yes, that's odd, but doesn't bar us from begining a comparison.

How fast have you documented the Xilinx cable going?

Reply to
cs_posting

yes that ODD

and yes it does prevent comparison, actually ;) the USB performance can be influenced by many things, it could be that amontec used FS only hub or root port as example,

well, lets wait up the "2.8 second howto" ..

Antti

Reply to
Antti

So how fast have _you_ gotten the xilinx cable to go?

You seem to be having more fun laughing at Larry's calendar challenges than actually seeking to compare performance.

The Amontec claimed performance, while yet unverified, doesn't seem unreasble to me, so I'm really curious if you have evidence that the xilinx cable is working faster than that for you?

Reply to
cs_posting

you should have realized that I want answer until I have actually compared the 1 cables in comparable environment.

I guess its all Larry's "marketing stuff" to promise things and then not deliver as promised. For me I see now reason or excuse to promise something "next monday" and then just ignore this. It sure is nice marketing, I have checked amontec website every few hours ;)

Antti

Reply to
Antti

actually as you have asked the same thing question SO MANY times, here is the answer YES, Xilinx Platform cable WORKS FASTER.

example: 11MBit bitstream, REDUCED TCK Clock to 12MHz, time : 2.547 seconds

On the test board the JTAG chain clock isnt optimal so I can not test at 24MHz TCK, I assume the speed performance would be noticeable.

this doesnt mean that Xilinx software and drivers are good, they are not, many JTAG operations could be carried out faster then do, but eh, this is the same thing as with Actel, they changed to use windriver USB drivers, and as result their programming times increased 2 times.

but hardware wise the Xilinx Platform USB cable is defenetly capable to get much better performance then any implementation of FT2232 in plastic box (== Amontec jtagkey, etc..) ever can. FT2232 has limitation on max JTAG clock of 6MHz.

Antti

Reply to
Antti

Dear Antti,

Sorry for the delay, but last Friday was the big CRASH. We received lightning on Amontec's House... the lightning comes in over LAN ! We lost 5 computers ! Our servers protected by UPS are safe, HOUFFF !

Strange meteo in Switzerland at this moment.

...

Reply to
Amontec, Larry

Hi all,

You may download the Amontec SVF Player from

formatting link

Already tested for programming Altera Lattice Xilinx FPGA s CPLD s and FLASH s. It can be use for programming AVR ATMEGA processors too. But you may use it as custom JTAG Boundary Scan.

Infini SCAN LENGTH ! Infini number of TAP (number of Targets) integrating Header and Trailer scans.

The amtsvfplayer.exe comes with c project source. Also, you may edit the source, customize it and re-compile a new SVF Player for your specific needs.

amtsvfplayer.exe -h to get help on usage.

You may execute SVF Files or SVF Lines. A SVF Line could be a concatenation of SVF commands.

You may adapt JTAG Frequency (FREQUENCY) via -frequencyFactor. In this way you do not need to edit the SVF yourself. When using

-frequencyFactor, the RUNTEST x TCK is automaticaly updated ...

Linux version ready to be published.

HAVE FUN WITH JTAG AND SVF ! ... but you need to have the JTAGkey.

Regards, Laurent

formatting link

Ann: via JTAGkey, you have a lot of ARM Debug Solutions too, including OpenOCD JTAG server, Crossworks, Yagarto ... !

Reply to
Amontec, Larry

Sorry to hear that.

You do appear to have recently posted some code for an SVF player. Always good to see manufacturer's providing real user flexibility in using their products!

Is this code the how-to for the JTAG speed claim?

Was the comparison to a Xilinx cable also run in SVF file mode, or did you have impact reading a native xilinx bitstream file?

Reply to
cs_posting

brrr :( too bad the lightning strike, and lost computers.

and too bad tha the "source code" published is 100% useless, as all the actual JTAG handling is hidden in an DLL and there is no source code available for it.

Antti

Reply to
Antti

Yes, this is the how-to code.

But the power of this solution is to be a generic JTAG solution because the SVF is a portable, FPGA / CPLD vendor independent! (Antti will not be OK with me, but vendor have to respect SVF specification. Our parser does :-) )

An other () : ( we will prove that we can program a ARM7 or ARM9 via SVF too !

Running a native xilinx bitstream could be a bit faster than we do not have to read and parse SVF file, but just take the .bit and upload it byte-after-byte. With the support of AmtXHAL, a native bitstream solution could be build in 1 or 2 hours for us (some hours for beginners) ... we could write an example if you need. Native bitstream stays custom solutions depending on number of chained Targets.

But you are right, in test / program production stage, secondes could be important.

Laurent

Reply to
Amontec, Larry

Dear Antti,

Did you try it ? or do you only search open sources ? Is a work too bad and 100% useless because it does not provide ALL in Open-Source? Do you make $$$ multi-donations for an Open-Source project? We have the replies.

Amontec team works daily for providing real user flexibility and solutions in using Amontec products ...

Our SVF player based on AmtXHAL is a very powerfull solution. And it is FREE. Our JTAG Hardware Abstration Layer is a big work too. We will put all code as Open-Source when our JTAGkey21 and JTAGkey24 will be ready for sales. These two new products will be based on USB2.0 high-speed processors and we already know that we will have the faster worldwide JTAG solutions).

If you are designing from AmtXHAL, the use of JTAGkey21 and JTAGkey24 will be transparent for you ! You have to see yourself, but you'll see soon !

Larry

Reply to
Amontec, Larry

Strongly disagree.

First you can use it as intended.

Then you can use the functions in the provided header file to accomplish various other jtag tasks.

And if you really want to understand it, well, you have a header file for Larry's DLL, and Larry's DLL calls the FTD2xx.dll. So you make up a fake version of the later, and see what a given trial call to Larry's DLL produces in terms of FTD2xx operations... Yeah, reverse engineering, but simpler than reverse engineering the xilinx stuff, and people have done that!

Reply to
cs_posting

eh there is absolutly no sense to RE Larry DLL's ;) its nothing magical to found there.

the "functions provided" did look like primitive replacement for something calles "command line parameter passing" - but well I only looked 2 minutes, maybe there is something more to see. But what I did see did look like useless. I would prefer just run from batch file, then using this customization API

Antti

Reply to
Antti

Electronic is not Magic but Logic. Only Physic is Magic! True Random Number Generator is Magic and Physic but use some Electronics!

It is very simple to talk and think about True Random Number but you need more than 2 minutes for developing a True Random Number Generator!

Bla - bla ... as your bla - bla Antti.

Antti, you CANNOT take 2 minutes and then resume by a "too bad" and by a "100% useless".

Laurent

Reply to
Amontec, Larry

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.