dre%ember@Sun.COM (David Emberson) (08/02/88)
Now that IBM has supposedly "invented" Harvard architecture with cache memories, I think it would be interesting to try to list machines that used this "new" technique prior to IBM's lawyers laying claim to it. I know that I worked on designs at Prime that had separate I and D caches as early as 1978. But none of these made it out of the lab, at least that I am aware of. How about the MC88000? Would it be a violation of IBM's "patent" to use one 88100 with *two* 88200s? I haven't seen the patent itself. Does anyone know what the patent claims are? If ever there was a case which proved the need for more vigilance on the part of the patent office and the legal profession, this is it. Unfortunately, abuse of the patent laws by big companies like IBM and the unfortunate legal- ization of the patenting of algorithms is going to have the effect of destroy- ing opportunities for small startup companies in the high technology sector. What investor would put even a small amount of seed capital into a company unless the poor entrepreneur had already gone to the expense of performing patent searches on even the most trivial of ideas to prove that the investment was safe? What garage entrepreneur can afford the $50K charged by your average shyster for EACH patent search? For the record, I do not speak for Sun Microsystems on this issue and I do not know anything other than what I have read in the press about the letter Sun received from IBM. McNealy says that we are not in violation of any of the patents, and I will take him at his word. At least he has integrity, unlike IBM. It seems to me that they feel threatened by new, better technology and want to buy time by turning loose their dogs, er, lawyers. Every scientist and engineer in this country should be appalled by their behaviour. Common subexpression elimination? Gimme a break! Single cycle mask and rotate instructions? Come on! Where will this end? We need a patent system that 1) prevents trivial claims, 2) is inexpensive, and 3) provides protection for small companies and entrepreneurs. Abuse of the system by greedy monopolists should be punished. Dave Emberson (dre@sun.com)
lamaster@ames.arc.nasa.gov (Hugh LaMaster) (08/02/88)
In article <62370@sun.uucp> dre%ember@Sun.COM (David Emberson) writes: > >Now that IBM has supposedly "invented" Harvard architecture with cache >memories, I think it would be interesting to try to list machines that used >this "new" technique prior to IBM's lawyers laying claim to it. I know that >I worked on designs at Prime that had separate I and D caches as early as >1978. But none of these made it out of the lab, at least that I am aware of. Of course, the CDC 6600 didn't use a D-cache, but it did have an I-cache back in 1964 that was separate from the processor-memory data path. I wonder if Cray patented this concept? (The cache on the 6600 was not quite randomly addressable, so does that make it not a cache?) It is perfectly clear that "RISC" is not a patentable concept, but there may be some real patents somewhere in that list. -- Hugh LaMaster, m/s 233-9, UUCP ames!lamaster NASA Ames Research Center ARPA lamaster@ames.arc.nasa.gov Moffett Field, CA 94035 Phone: (415)694-6117
baum@Apple.COM (Allen J. Baum) (08/03/88)
[] >In article <62370@sun.uucp> dre%ember@Sun.COM (David Emberson) writes: > >Now that IBM has supposedly "invented" Harvard architecture with cache >memories, I think it would be interesting to try to list machines that used >this "new" technique prior to IBM's lawyers laying claim to it. The IBM patent itself has a list of prior art, together with explanations of the differences. You have to be very careful about reading just exactly what they are claiming. Claim No.1 says basically: Separate I & D caches, with tag stores, and valid bits settable under program or system control. The relevant differences are the "settable under program control" part, i.e. they have cache control instructions. To continue: Claim 2: #1+Both caches can be accessed simultaneously, but cache<->main mem. path only allows one at a time. Claim 3: #1+Dirty bit, settable under program or system control (more insts.!) Claim 4: #3+Data cache can write back to mem, but not Icache. Claim 5: #4+means for changing the valid bit checks the Icache tags Claim 6: #5+Forward data directly to CPU on miss (as well as sending to cache) Claim 7: #5+Means for inhibiting the setting of the dirty bits (more insts.!) Claim 8: #7+2 way set assoc. with LRU bit to determine which set Claim 9: #8+bit is set to indicate LRU on hit or miss Claim 10:#9+don't update tags if hit occurs, but valid bit is off. Claim 11:#10+tag store is faster than data store. Claim 12: 2 Caches w/tag+data sections, valid bit/line, settable by sys/prog, >=1 dirty bit/line, settable by sys/prog, no write path from Icache to main mem Claim 13:#12+Forward data directly to CPU on miss (as well as to cache) Claim 14:#13+means for preventing setting of the dirty bit Claim 15:#14+ each cache is 2 way set assoc. w/LRU Claim 16:#15+don't update tags if hit occurs, but valid bit is off. I may have missed the emphasis of the claims in paraphrasing them, but I believe the thrust is that there is no hardware to ensure that writes to the DCache get reflected in the Icache, and there are cache control instructions. The disclosure lists 6 cache control instructions: Set Data Cache Line: write tag, but don't actually fetch data; this line will used for temp storage, and you don't care what prev. data was. Invalidate Data Cache Line: force susequent accesses to go to main mem. Invalidate Inst Cache Line: ditto Store Data Cache Line: make sure main mem has valid copy Load Data Cache Line: prefetch? don't forward to CPU! Store & Synch. Data Cache Line: make sure main mem has valid copy, don't start new inst. until this completes In summary: IBM patents are not without foundation. They are diligent about researching prior art. The patents are strong, well written patents, and do not make wild claims about inventing fundamental concepts (unless, of course, they actually did, which is not an uncommon case). -- {decwrl,hplabs,ihnp4}!nsc!apple!baum (408)973-3385
dre%ember@Sun.COM (David Emberson) (08/03/88)
In article <15086@apple.Apple.COM>, baum@Apple.COM (Allen J. Baum) writes: > > In summary: IBM patents are not without foundation. They are diligent about > researching prior art. The patents are strong, well written patents, and do > not make wild claims about inventing fundamental concepts (unless, of course, > they actually did, which is not an uncommon case). It seems to me that in this case they patented, among other things, setting a bit under software control. The idea of managing the contents of a cache memory under software control sounds like something that is a fundamental, obvious idea to me. I have no doubt that even this patent is well written. IBM has the money to afford the best legal talent in the country. But the garage entrepreneurs don't, and that is my major point. The present system stifles innovation and works to the benefit of the largest, most powerful, richest companies. The patent system was intended to protect inventors, not monopolies.
oconnor@nuke.steinmetz (Dennis M. O'Connor) (08/03/88)
I don't know when IBM filed their patent application, but
as of October 1968, the following were published "prior art"
( We at GE published them as a goverment contract final report )
An article by baum@apple.UUCP (Allen Baum) says:
] The IBM patent itself has a list of prior art, together with explanations of
] the differences. You have to be very careful about reading just exactly
] what they are claiming. Claim No.1 says basically: Separate I & D caches,
] with tag stores, and valid bits settable under program or system control.
] The relevant differences are the "settable under program control" part,
] i.e. they have cache control instructions. To continue:
The RPM40 I-Cache has cache control instructions, including loading
and storing tags and their valid bits. ( If you're going to
software manage the cache, you have to be able to save and restore
it's context, at least the part you manage, on task switches. )
] Claim 2: #1+Both caches can be accessed simultaneously, but
] cache<->main mem. path only allows one at a time.
This was part of a proposal GE submitted to the Government in late '86
] Claim 4: #3+Data cache can write back to mem, but not Icache.
This was part of the final report.
] Claim 6: #5+Forward data directly to CPU on miss as well as sending to cache
This is exactly how our I-Cache works, and is part of the October '86 report.
] Claim 11:#10+tag store is faster than data store.
No kidding. Is this new ? It wasn't to us in October of 1986.
] Claim 13:#12+Forward data directly to CPU on miss (as well as to cache)
Hey I've seen this claim before ! This was part of the Oct.'86 final report.
] I may have missed the emphasis of the claims in paraphrasing them, but I
] believe the thrust is that there is no hardware to ensure that writes to the
] DCache get reflected in the Icache, and there are cache control instructions.
"cache control instructions" were part of our Icache in the final report.
] In summary: IBM patents are not without foundation. They are diligent about
] researching prior art. The patents are strong, well written patents, and do
] not make wild claims about inventing fundamental concepts (unless, of course,
] they actually did, which is not an uncommon case).
"prior art" is almost impossible to be current on, in a field moving
as fast as computer architecture. The date of filing of the IBM
patents is very important. When were they filed ?
] {decwrl,hplabs,ihnp4}!nsc!apple!baum (408)973-3385
--
Dennis O'Connor oconnor%sungod@steinmetz.UUCP ARPA: OCONNORDM@ge-crd.arpa
"Never confuse USENET with something that matters, like PIZZA."baum@Apple.COM (Allen J. Baum) (08/04/88)
[] >In article <62555@sun.uucp> dre%ember@Sun.COM (David Emberson) writes: >In article <15086@apple.Apple.COM>, baum@Apple.COM (Allen J. Baum) writes: >> >> In summary: IBM patents are not without foundation...blah blah blah >It seems to me that in this case they patented, among other things, setting a >bit under software control. You are misreading my paraphrasing. They did not patent setting a bit under software control. They patented setting a cache-line-dirty, or cache-line-valid bit under software control in a system with split I & D caches. There is a world of difference in that. > >The idea of managing the contents of a cache memory under software control >sounds like something that is a fundamental, obvious idea to me. Its very easy to say this with the benefit of hindsight. But, at the time this was invented, it was neither fundamental, nor obvious. After all, no one else had done it, and it was taught that caches were these invisible things that sped up your program, and programmers couldn't tell they were there. All of a sudden, someone (presumably) at IBM said "Lets turn it on its head- LET the programmer know about the cache, and even be able to control it!" These kinds of intuition are not obvious. They're clever, and patentable. You can argue whether IBM was in fact the first to come up with the idea. If they weren't, you can argue whether the particular combination of features that they have claimed (and you have to read the patent VERY carefully to catch exactly what they are and are not claiming), but that's about all you can argue about (on the basis of patent law). > I have no >doubt that even this patent is well written. IBM has the money to afford the >best legal talent in the country. But the garage entrepreneurs don't, and >that is my major point. The present system stifles innovation and works to >the benefit of the largest, most powerful, richest companies. The patent >system was intended to protect inventors, not monopolies. It is, unfortunately, a fact of life that big companies have more resources that they can devote to coming up with new ways of doing things. Patent law does not penalize big companies because they have this advantage. It does not completely stifle innovation, or small companies, or I wouldn't be working here! These ideas would be patentable even if a garage shop came up with them. They are not patentable merely because a big company paid smart lawyers lots of money to write the patent disclosure. -- {decwrl,hplabs,ihnp4}!nsc!apple!baum (408)973-3385
hankd@pur-ee.UUCP (Hank Dietz) (08/04/88)
In article <62555@sun.uucp>, dre%ember@Sun.COM (David Emberson) writes: > In article <15086@apple.Apple.COM>, baum@Apple.COM (Allen J. Baum) writes: > > In summary: IBM patents are not without foundation. They are diligent about > > researching prior art. The patents are strong, well written patents, and do > > not make wild claims about inventing fundamental concepts (unless, of course, > > they actually did, which is not an uncommon case). ... > The idea of managing the contents of a cache memory under software control > sounds like something that is a fundamental, obvious idea to me. Well, complete software control IS new, but it isn't neatly patentable and IBM isn't doing that -- Chi-Hung Chi and I are and we've been publishing on it for a year and a half. Further, the crude control IBM has patented seems to me seems SIMILAR to the control used in many machines and proposals. However, I read IBM's patent (as summarized by Allen Baum) as claiming specific cache control instructions: not claiming the general, abstract, concept of software cache control. ... > The patent system was intended to protect inventors, not monopolies. No. The patent system was designed to protect investment in technological development, hence encouraging development of innovative new devices. An invention is not supposed to be patented UNLESS you will market it -- this is the logic behind patent-maintenance fees. The unfortunate thing about the IBM patents is not the patents themselves, but the PUBLIC USE OF THEM AS A SCARE TACTIC. As far as I can tell, the patents are really quite weak -- they protect very little and, although very carefully written, are easily challenged or circumvented. But who would dare challenge IBM? Worse still, who is large enough to survive the legal costs of being challenged (rightly or not) by IBM? It's just my opinion, but I think IBM stepped way over the line in saying that these patents were to be used to "reclaim" their inventions which other companies have stolen. However, it's a moral line -- not a legal line -- that they've crossed. -hankd
mslater@cup.portal.com (08/04/88)
We have made copies of the 16 IBM patents in the the list posted by Alan Baum, and will send a set to anyone for $35 (which basically just covers our costs). It's a bit over 300 pages. Not the most entertaining reading, but if you want to understand what they claim, this is the source. (If you do understand what they claim, please give me a call :-).) The fine print: add $20 if you're outside the US and Canada; we accept MC and VISA. Michael Slater, Editor and Publisher Microprocesor Report, 550 California Ave., Suite 320, Palo Alto, CA 94306 415/494-2677 mslater@cup.portal.com sun!portal!cup.portal.com!mslater
bpendlet@esunix.UUCP (Bob Pendleton) (08/04/88)
From article <62555@sun.uucp>, by dre%ember@Sun.COM (David Emberson): > The idea of managing the contents of a cache memory under software control > sounds like something that is a fundamental, obvious idea to me. I have no > doubt that even this patent is well written. Yes, I can even remember discussing the idea in a undergraduate class. We came to the conclusion that it wasn't a very good idea. This was 12 to 14 years ago. > IBM has the money to afford the > best legal talent in the country. But the garage entrepreneurs don't, and > that is my major point. The present system stifles innovation and works to > the benefit of the largest, most powerful, richest companies. Yes, and to the benefit of small organizations and individuals with truely new ideas. > The patent system was intended to protect inventors, not monopolies. This statement contradicts my basic understanding of the patent system. As I understand it, the idea was to give the inventor a limited monopoly on the use of a technique, plant or animal, in exchange for releasing a complete description of the technique... The goal is to stimulate innovation and protect society from loss of knowledge by giving the inventors a reason to publish information they might otherwise wish to keep as trade secrets. It is my understanding that the patent system was intended to benefit society by granting monopolies to inventors. I've been told, by a patent lawyer, that the fact that patents are very difficult to enforce has PREVENTED large investments in research because you can't count on making any money from it. Bob P. -- Bob Pendleton @ Evans & Sutherland UUCP Address: {decvax,ucbvax,allegra}!decwrl!esunix!bpendlet Alternate: utah-cs!esunix!bpendlet I am solely responsible for what I say.
baum@Apple.COM (Allen J. Baum) (08/05/88)
[] >In article <11717@steinmetz.ge.com> oconnor%sungod@steinmetz.UUCP writes: >I don't know when IBM filed their patent application, but >as of October 1968, the following were published "prior art" >( We at GE published them as a goverment contract final report ) I will blithely assume that 1968 is a typo for 1986.... >"prior art" is almost impossible to be current on, in a field moving >as fast as computer architecture. The date of filing of the IBM >patents is very important. When were they filed ? The patent was filed in May,1982. The actual work was done prior to that, of course, so that any prior art would have to prove that the idea came up before it was documented in the lab notebooks at IBM, which was prior to May, 1982. -- {decwrl,hplabs,ihnp4}!nsc!apple!baum (408)973-3385
fotland@hpihoah.HP.COM (Dave Fotland) (08/05/88)
IBM's RISC patents are probably based on the 801 work that happened in the late 70's, way before the Oct 86 final report. In the US it is the date of invention that has priority in patents, not the date of filing. Since IBM disclosed some of this stuff back in the early 80's I'm sure they filed back then also (since you only have a year from the date of disclosure in US).
urjlew@ecsvax.uncecs.edu (Rostyk Lewyckyj) (08/05/88)
Mr. Baum claims that Ppabiggness does not provide an advantage
in patenting and patents. He also seems to feel that because
IBM managed to get a patent on a precise implementation of
a perhaps common art technique, it must have been an
innovation.
Well I disagree with thatose opinions.
Big companies have staff lawyers to put together the paperwork
to claim even small "innovations", and they do this almost as
a matter of course. A small complany ofr individual will not
often do this. First because he may think the innovation too
trivial. Second because of the expense and the knowledge that
he can not defend the patent if it is infringed.
As regards to sothe specifics of "setting the dirty cache bit
under software contorlrol in a split I and D cache machine".
Ever since there were different memory hierarchies there
have been caches. (Read a matrix in from tape, then write it
out. etc). Initially these theings were very much under an
individual proggrammers control. (eg getting Swords from backing
store on a CDC). Then some people worked very hard to try and
do this kind of fetching of data to the faster storage (cache)
automatic and transparent to the user. I have heard that IBM
specifically Dr. Blauwe (sp?) made this possible on the 360s
with the Blauwe box (ref. Computer architecture class taught
by Dr. Fred Brooks). So it is not surprising that people were
mainly thinking of caches as transparent to the programmer.
HOwever the logic to check on the validity of the cache had
to be there in a harware program. And allowing programmer
control of the dirty bit flag is simply going back to when
the programmer did it all himself. What IBM did is implement
a specific instance of something that needs to be done to make
caching work, and happened to patent it because they try to
patent everything. NOw they are trying to wield the patent
as a club.
-----------------------------------------------
Reply-To: Rostyslaw Jarema Lewyckyj
urjlew@ecsvax.UUCP , urjlew@tucc.bitnet
or urjlew@tucc.tucc.edu (ARPA,SURA,NSF etc. internet)
tel. (919)-962-9107jesup@cbmvax.UUCP (Randell Jesup) (08/05/88)
In article <15129@apple.Apple.COM> baum@apple.UUCP (Allen Baum) writes: >It is, unfortunately, a fact of life that big companies have more resources >that they can devote to coming up with new ways of doing things. Patent law >does not penalize big companies because they have this advantage. It does not >completely stifle innovation, or small companies, or I wouldn't be working >here! These ideas would be patentable even if a garage shop came up with them. >They are not patentable merely because a big company paid smart lawyers lots >of money to write the patent disclosure. Apple's success hasn't depended very much on patents. The way patents work nowadays does greatly benefit large companies. It is not uncommon for an individual or small company to be granted a patent, than have the big companies either use it without paying, or to pay their lawyers to "break" the patent, using their much bigger resources to try to prove the patent shouldn't have been granted. The patent searches done by the patent office often miss things: this causes much more grief for individuals than big companies, since the individual assumes that since he got the patent, it can't be taken away. If an individual tries to sue even medium sized companies that have used his patent without paying, it can take 10 or more years to get compensation, if ever. Many people with patents have been driven into bankruptcy trying to defend their patent (and few law firms will defend patents on contingency). Of course, this isn't limited to just patents, though it may be worst there. Copyright law also suffers from "justice is for those with the most money" tendency of our civil laws. Witness the Lotus, Apple, and a number of other cases (and those are just the most visible). Lets not go too deep into this, or if you want to, move discussion to comp.misc. The whole copyright thing has been hashed out too many times. Let's stick to patents, or better yet (gasp) architectures. -- Randell Jesup, Commodore Engineering {uunet|rutgers|allegra}!cbmvax!jesup
alvitar@madhat.UUCP (Phil Harbison) (08/07/88)
In article <62370@sun.uucp>, dre%ember@Sun.COM (David Emberson) writes: > ^^^ > Now that IBM has supposedly "invented" Harvard architecture with cache > memories ... > > If ever there was a case which proved the need for more vigilance on the > part of the patent office and the legal profession, this is it.... > > Common subexpression elimination? Gimme a break! Single cycle mask and > rotate instructions? Come on! Where will this end? > > We need a patent system that 1) prevents trivial claims, 2) is inexpensive, > and 3) provides protection for small companies and entrepreneurs. > Is it just me or do others see the irony in this statement? Here we have a guy complaining about IBM's ability to patent such "obvious" things as separate I&D caches, common subexpression elemination, etc. However, a quick glance at the authors signature line tells me he works for SUN Microsystems, the same company which was able to patent the "revolutionary" concept of using the least significant bits of an ad- dress as the row addresses for a bank of DRAMs. Is this not a case of the pot calling the kettle black? Of course I am referring to the patent SUN has on using the bits which bypass the MMU as row addresses, thereby allowing the row access to overlap with the MMU translation time. Never mind the fact that it is convention to use the LSBs as the row address anyway, and that the over- lap potential should be obvious to anyone skilled in digital circuit design. In my opinion, SUN is just a little IBM that just happens to have better tastes in processors and operating systems. :-) By the way, this is not meant as a personal attack on Mr. Emberson. My complaints are with the SUN coorporation, not its individual employees, who for the most part are very nice people. ---- Live: Phil Harbison USPS: 3409 Grassfort Drive, Huntsville, AL 35805-5421 UUCP: madhat!alvitar PSTN: 205-881-4317 or 205-830-4310x210
eugene@eos.UUCP (Eugene Miya) (08/10/88)
In article <8729@pur-ee.UUCP> hankd@pur-ee.UUCP (Hank Dietz) writes: >The unfortunate thing about the IBM patents is not the patents themselves, >but the PUBLIC USE OF THEM AS A SCARE TACTIC. It was said during the "auto crisis:" when the new regulations come out, Detroit hires 200 lawyers, XXX (back then it was the Japanese), hires 200 engineers. Another gross generalization from --eugene miya, NASA Ames Research Center, eugene@aurora.arc.nasa.gov resident cynic at the Rock of Ages Home for Retired Hackers: "Mailers?! HA!", "If my mail does not reach you, please accept my apology." {uunet,hplabs,ncar,decwrl,allegra,tektronix}!ames!aurora!eugene "Send mail, avoid follow-ups. If enough, I'll summarize."
baum@Apple.COM (Allen J. Baum) (08/11/88)
[] >In article <5536@ecsvax.uncecs.edu> urjlew@ecsvax.uncecs.edu (Rostyk Lewyckyj) writes: > >Mr. Baum claims that biggness does not provide an advantage >in patenting and patents. He also seems to feel that because >IBM managed to get a patent on a precise implementation of >a perhaps common art technique, it must have been an >innovation. Well, I don't think it was common art, not in 1982. >Big companies have staff lawyers to put together the paperwork >to claim even small "innovations", and they do this almost as >a matter of course. IBM does not patent every innovation. Anyone who has seen the IBM Technical Disclosure bulletin would understand that immediately. The Bulletin comes out once or twice a month, is a couple of inches thick and full of the ideas that they think are clever, but not worth their time to patent. They are published so someone else can't patent them, and force IBM to pay royalties. When I worked for HP, we did not patent every little innovation. The engineers would write a disclosure, and HP would go ahead with only those patents that they thought were worth it. Lots of patents are too much trouble even for a big company with lots of lawyers. -- {decwrl,hplabs,ihnp4}!nsc!apple!baum (408)973-3385
ok@quintus.uucp (Richard A. O'Keefe) (08/11/88)
In article <15353@apple.Apple.COM> baum@apple.UUCP (Allen Baum) writes: >IBM does not patent every innovation. Anyone who has seen the IBM Technical >Disclosure bulletin would understand that immediately. The Bulletin comes >out once or twice a month, is a couple of inches thick and full of the ideas >that they think are clever, but not worth their time to patent. Sounds great. (I hope they put software ideas there too!) What does it cost, and how do I get a sample issue?
baum@Apple.COM (Allen J. Baum) (08/12/88)
[] >In article <275@quintus.UUCP> ok@quintus.UUCP (Richard A. O'Keefe) writes: >> (Allen Baum) writes: ...the IBM Technical Disclosure Bulletin comes >>out once or twice a month, is a couple of inches thick and full of the ideas >>that they think are clever, but not worth their time to patent. > >Sounds great. (I hope they put software ideas there too!) >What does it cost, and how do I get a sample issue? I'm not sure how you get hold of it. I've been trying to find out for a couple of years. Unlike other IBM journals, it might only be available to libraries and educational institutions. Considering the amount of shelf space a years worth of it would take, its probably not cheap. If anyone knows more details, PLEASE post! -- {decwrl,hplabs,ihnp4}!nsc!apple!baum (408)973-3385