Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Microcap & Penny Stocks : Patriot Scientific - PTSC -- Ignore unavailable to you. Want to Upgrade?

To: bob who wrote (4923)	6/3/1998 10:23:00 AM
From: sdr	Read Replies (1) \| Respond to of 8581

Go CAP !!

To: bob who wrote (4923)	6/3/1998 10:24:00 AM
From: R Sandvig	Respond to of 8581

Hey, we have NEWS! Aint it great! We are all going to be RICH! Rich S.

To: bob who wrote (4923)	6/4/1998 12:06:00 PM
From: Urlman	Respond to of 8581

Yakity Yak over @ Usenet (BTW I also forgotto mention UniView is also working with Motorola) REPOSTED FROM USENET: Subject: Re: Benchmarking Java: But how does it compare to C/C++? From: msimon@tefbbs.comDate: 1998/06/04 Message-ID: <35749bcd.8112092@news.megsinet.net> Newsgroups: comp.lang.java.machine[More Headers] [Subscribe to comp.lang.java.machine] Haven't done your homework I see. If each opcode is 1 byte and it is a 32 bit machine each memory cycle brings in 4 instructions. That would give a 66 MHz machine. If the op-codes are 5 bits (iTV machine, MuP21, F21) that is 100MHzperformance. Or you could design a slow machine - ie 66MHz clock that can do up to 5 things per instruction. Average 2 operations per clock and you get 120MHz performance. Peak 333MIPS - avg probably around 60MIPS. You ought to read 'Stack Computers' and some of the other sites I have mentioned. Zero operand machines are VERY fast for the amount of invested technology.SimonMarty Fouts <usenet-user@usa.net> wrote: > >> In article <3572b97f.542604@news.megsinet.net>, > >> msimon@tefbbs.com proclaimed:>> [snip]> > > A core representation need only have 4 registers - Top of Data > > Stack, Next on Data Stack, Top of Return Stack, and Program > > Counter. Machines like this have been built in the 5- 10K gate > > region for years. Execution speed averages 1 instruction per clock > > tick. These are high level instructions> >Stack machines at current memory cycle times require massive >caches. You aren't going to get anywhere near 1 ipc at 100mhz as you >wait patiently for a 60ns memory part.> > > Patriot Scientific has a Java Machine called the SHboom. Designed > > by Chuck Moore of FORTH fame. Not quite in the price range I > > mentioned but headed in that direction (they have to get the > > volumes up). (Patriot Scientific says that 38% of Java Byte Codes > > are machine instructions).>>IE, another $300/chip in 10k quantity part? I think small quantities are in the $10 to $40 range. But I could bemistaken. If you want to be assured of a supply you could do a core on a Xilinx chip for $50 - $100 depending on speed (1ea price) These kinds of processors are not very hard to design.>> [snip]> >So none of these things have anything to do with JavaEngine, and by >the time I build the glue logic around them the real cost will be in >the 300 dollar range as well. In one each volumes yes. Sun's Engine on a board is on the order of $350 for volumes of 10K. Meaning about $750 to 1K in ones. > > As to the FPU - simulate in software (unless you are number > > chrunching). Real People use fixed point because it is > > cheaper. Floating point is a crutch for programmers who don't want > > to think about numbers.>> >So the key word I should have recognized was _imagine_, then?> There is at least one low cost FORTH processor on the market dedicated to Java and several more that can be easily adapted. Not too shabby for a language only a few years old. You know of any processors that execute C bytecodes? With the small gate counts required and FPGAs available for experimentation I expect to see an explosion of Java processors in thenext year. I mean 2 stacks, 2 registers, an alu and some glue. How hard can itbe? With your understanding I'm sure you could turn one out in a few months of part time effort.Simon Opinions expressed herein are solely my own and may or may not reflect my opinion at this particular time or any other.

To: bob who wrote (4923)	6/4/1998 3:53:00 PM
From: FastC6	Read Replies (2) \| Respond to of 8581

Cap, How much do you think it will cost a company to make their own java processor from scratch including costs from R&D to final product? SS

To: bob who wrote (4923)	6/4/1998 9:16:00 PM
From: Urlman	Respond to of 8581

Sun Micro Loses ANOTHER Java Processor Executive: Computergram International June 3, 1998 CHORDIANT NAMES FORMER SUN EXEC AS CEO ÿÿÿChordiant Software Inc, Palo Alto, California-based provider of call center software, has appointed Sam Spadafora its president and chief executive officer. Spadafora succeeds Carol Realini who will continue as a working chairman of the board. He comes to Chordiant after spending the past four years at Sun Microsystems Inc as vice president of worldwide field operations for the microelectronics business. At Sun Spadafora was responsible for the company's SPARC, Java processor and Java embedded consumer product lines. Prior to his tour of duty at Sun, Spadafora had held a variety of executive level positions in general management, sales and marketing over the past thirty years at companies including The Santa Cruz Operation Inc, Altos Computer Inc and Xerox Corp.