- cypherpunks-legacy - lists.cpunks.org

DES, MMX, and FPGAs
by Trei, Peter 17 Dec '03

17 Dec '03

In DES key search, the speed of a given machine is controlled by three main factors: clock speed, the algorithm used, and the availability of multiple issue (the last refers to perfoming 2 or more instructions in parallel). Clock speed can be easily factored out for comparison purposes. Algorithm is harder; a bitslice algorithm is a clear win on 64 bit machines with lots of registers, but is roughly equivalent to non-bitslice algorithms on 32 bit machines with a low register count (eg, Intel processors, even using MMX). Multiple issue can greatly speed up a specificly tuned assembly language version. I can't speak for the non-intel processors below, but the Pentium code below is all tuned for the vanilla Pentium, and does not get a boost from the PPro or P2. Let's do the numbers: d.net peak keys per second: 34,430,460,000 cpk (clocks per key) = (equiv * clock speed)/34B Processor clock (MHz) equiv cpk DEC Alpha 321064 142 (4) DEC Alpha 21066 220 (4) DEC Alpha 21064 533 11,264 176 (1) DEC Alpha 21164 (Deschall) 94 (3) Sun Ultra I 167 15,316 75 (1) SGI r10k 69 (4) Macintosh PowerPC 604e 200 68,859 403 (1)(2) Macintosh PowerPC 604e 200 137 (4) Intel Pentium II 333 22,393 219 (1)(6) Intel Pentium Pro 232 (4)(6) Intel Pentium 166 41,712 203 (1) Intal Pentium (Bryddes) 195 (5) Intel 486DX2 66 399,374 775 (1) Intel 386SX 20 7,446,033 4380 (1) (1) D.net press release. (2) This figure comes from the press release. I don't believe it for a moment. The next line gives a speed determined from a posting to the d.net mailing list, and is much more believable - if anything, it appears a little slow. (3) Deschall client (4) based on d.net mailing list posting. (5) Sven Mikkelsen's page. (6) I'm not aware of any fielded bitslice software for Intel MMX platforms. I suspect that it would not speed things up too much - there are simply too few registers, and you'd spend a lot of time moving stuff in and out of cache. I wouldn't be suprised if some of the figures above are incorrect, but it's clear that the best processors are getting down to 70-100 clock cycles per key. Wiener's keysearch engine unrolls the 16 rounds of DES, and is thus able to run in a similar way to an assembly line; once it's 'filled up' it tests 1 key per clock cycle. Any other version would run a lot slower. Some highly uninformed blather about FPGAs follows: Wiener's paper uses custom VLSI chips, and claims a need for 26k "gates" or 78257 "sites". I'm not sure how this translates to silicon requirements in FPGAs - it seems that "gate" and "site" are nominal terms which are not directly equivalent to actual, physical gates. I've found one reference to a general DES FPGA implementation which requires 25k gates, though it's not clear if this reuses the round circuitry, or 'unrolls' it (I suspect the latter, which would be good). I'm also not certain if a FPGA running at, for example, 50 MHz, can actually run logic operations at that speed. Some things I've read suggest that FPGAs run ops at several clocks per operation - which is still a lot better than a general purpose microprocessor. The fastest and largest FPGAs I'm aware of are the Xilinx 4000 series: The xc40125 has about 80,000 gates, and runs at around 100 MHz (and costs $1500!!). Suppose we use this chip, and everything works out optimally. We can fit 3 des engines per chip, and get 300 Mkeys/sec checked. To duplicate d.net's speed, we would need about 115 chips, or about $170k for the chips alone. A better choice would be the Xilinx Spartan XL series. This might be in the $20 - $40 a chip, and would run at 80MHz, and would fit only one des engine. This would be around $9k - $17k, which is still pretty high for an individual. At this speed (34Gkey/sec), it would still take 12 days to search half the keyspace, and the prices quoted are for the chip alone, no boards, chassies, or SW, and the speeds are assumed to be 1 key/engine/clock, which is optomistic. On the plus side, implementing a DES cracker in an FPGA actually affords us some efficiency savings over a non-reconfigurable chip. The cipher and plaintexts can be compiled into the design, and if you're willing to reload the chip frequently, the slower-changing portions of the key can be built in as well. This cuts the gate count considerably. I'd really like to see the next DES challenge taken on by some organization with a few hundred under utilized FPGA boards. Peter Trei ptrei(a)securitydynamics.com

1 0

They're running scared.
by jim bell 17 Dec '03

17 Dec '03

Jim Hoagland's newspaper column from April 4, 1996: Governments awakening to threat of the Internet The computer and modem now downsize the globe, enabling citizens to vault over walls of secrecy, law and control erected by governments. Still gathering steam, the information revolution is creating a new generation of ticklish foreign policy and national security problems for the world's governments. They are organized to operate in a heirarchical world of borders and customs posts and to keep out the unwanted, the unhealthy or the dangerous. But the boundaries of cyberspace are unfixed and amorphous. They are being determined more by the availability and cost of communication modems, sophisticated software, satelite stations, encryption techniques and other data processing technology than by government fiat. An example of cyberspace's potential for harm surfaced last week when France asked the United States to crack down on a San Diego-based Islamic group that posts instructions on the Internet for assembling inexpensive bombs like those exploded on the Paris subways last year. French officials traveling or posted abroad fear they are the intended targets of these homemade bombs, the Quai D'Orsay's senior Middle East expert, Denis Bouchard, told American diplomats at a meeting last week on international terrorism in Washington. State Department officials offered the French sympathy. But they did not hold out much hope they coudl act on the sparse information the French provided. The line between computer-driven incitement to terrorism and electronic free speech still has to be drawn in the brave new cyber world. The inchoate nature of that world was underscored by the disclosure March 29 that U.S. authorities had charged an Argentine student with three felonies for illegally entering Pentagon and other U.S. military computers to obtaining confidential files on satellites, radiation and energy-related engineering. But Julio Cesar Ardita, 22, who raided Washington files from his home in Buenos Aires, cannot be extradited under American-Argentine treaties, which do not cover these alleged national security violations. Governments are waking late to the implications of individuals and small groups operating across boundaries and oceans to bypass, introde upon or flip and electronic finger at bureaucracies that have controlled or regulated the security and business of nations for centuries. The implications are particularly dramatic for totalitarian regimes that brook no open dissent. China seeks to impose a government monopoly over economic data transmission into China to go along with the draconian political censorship already practiced on the nation's traditional media. But as long as the Middle Kingdom remains part of the International telephone system with its faxing and modem capabilities, words and facts the communist leadership abhors will spread faster than Big Brother can track them. The world stands roughly where it stood as television began to reshape politics, and policy-making, in ways that we still do not fully understand. A new communication technology arrives to change what we think, as well as how we think and communicate. Traditionalists fear anarchy (or obsolescence). Optimists foresee the best of all worlds, with Orwell's 1984 predictions of Big Brother tracking and brainwashing everyone through television proven to have been 180 degrees off course. But the picture is in fact mixed. Governments have begun to talk seriously to each other about controlling the computer revolution. The Pentagon is studying the information highway as the route to complete domination of the battlefield and thus the ultimate source of power. THe FBI, IRS, and CIA are determined to keep you from being able to encode and transmit information they want to see. Orwell may turn out to have been premature, but not wrong. The struggle over the course of the information revolution is only beginning. The bureaucracies that are most threatened still have powerful hands to play. There is no guarantee that cyberspace will provide the world with the era of new freedoms that now seem likely. That battle is still to be fought, and won. [end of article] Articles such as this are interesting because they appear to be written without any illusion that the interests of governments are anything other than just that, interests of governments. They are NOT the interests of the average citizen.

1 0

Re: So, what crypto legislation (if any) is necessary?
by jim bell 17 Dec '03

17 Dec '03

The following is a comment of mine that Unicorn didn't respond to. My comments are included as well. I repeat this because I just saw a quote from Lysander Spooner which was appropriate under the circumstances. Those who read Unicorn's notes will notice that he frequently ignores embarrassing gaffes that he makes. [beginning of re-quote] At 08:56 PM 3/30/96 -0500, Black Unicorn wrote: >> Naturally, you won't address this >> problem, but the man-on-the-street is more realistic about his own privacy. >> How many times must I raise this issue? How many times do you ignore it? >> Face it, people are smarter than you give them credit for. They will simply >> not tolerate any more shit from the government. > >Funny, the latest primary has been one of the highest voter turn outs in >quite a while (except in Deleware). Considering those are the law-and-order >types who are most likely to invade personal liberities, I think its a >bit hard to make the case that the temper of the country is anything but >very pro-political process. Unicorn again displays his cluelessness. If people know that the system is sick, and they believe (even wrongly) that the only way to fix the problem is through "the political process," they can reasonably be expected to take one last, desperate effort at fixing the situation. That doesn't make anybody "pro-political-process," in fact they could be disgusted with the lack of progress that this system produces. They simply believe that they have no alternative. To put it in simple terms that even you should be able to understand, the fact that a drowning person moves his arms and legs around a lot doesn't mean that he LIKES to swim, it may merely mean that he likes drowning even less. [end of re-quote] Spooner's quote follows: "Doubtless the most miserable of men, under the most oppressive government in the world, if allowed the ballot, would use it, if they could see any chance of thereby meliorating their condition. But it would not, therefore, be a legitimate inference that the government itself, that crushes them, was one which they had voluntarily set up, or even consented to." Lysander Spooner

2 1

Congressbastards getting a clue
by David Honig 17 Dec '03

17 Dec '03

http://allpolitics.com/1998/03/10/ap/mcdermott.lawsuit/ Summary: History: Florida couple taped some Congressmen (including Gingrich) talking on their cellphones. Gives the tape to another Congressman A, who passes it on to the press. Florida couple has been convicted. News: Now one of the taped Congressmen is suing the Congressman who leaked the tape. When will they learn? Mathematics >> law ------------------------------------------------------------ David Honig Orbit Technology honig(a)otc.net Intaanetto Jigyoubu "I do believe that where there is a choice only between cowardice and violence, I would advise violence." Mahatma Gandhi

1 0

GoldMoney
by John Young 17 Dec '03

17 Dec '03

http://www.goldmoney.com/home.html Essays on the future of money, electronic, digital and "GoldMoney," based on US Patent No. 5,671,364 One or so cypherpunks referenced.

1 0

Re: the cost of untracability? Re: the cost of untracability?
by mixmaster＠remail.obscura.com 17 Dec '03

17 Dec '03

[Please excuse the minor taxpunks diversion; there's a relevant point here, aside from the fact that we're approaching April 15 and discussing technologies that will let some people not care if they're approaching April 15 in a few years :-] At 09:28 AM 4/7/96 -0800, Jim Bell wrote: >FWIW, I think that there is no capital-gains-type tax on currency >conversions. In other words, if I take dollars and buy yen today, >and the interconvert rate changes and I convert back and make a >"profit," that is not considered income. What it's worth is fairly minimal, i.e. the IRS doesn't see it that way. If you did transaction Y and then transaction X and now have $100 more than you did before you started, they think that's capital gain, whether you were buying and selling Yen, soybeans, or mutual funds. The only difference if you're buying and selling ecash from Bank Foo is that it may be easier to not tell them about the transaction if it were all encrypted and outside your home jurisdiction. What they don't know won't hurt them... On the other hand, depending on what country your bank is in, there may be taxes or fees or bribes charged by the bank's home country, which would get passed along to you either directly or indirectly. > And most of the "interest" will simply be the avoided inflation loss > that would have otherwise occurred. Buying ecash may be equivalent > to buying an absolutely non-inflating currency that the government > can't manipulate. No, it's equivalent to buying private-bank currency, which may be denominated in dollars or ECUs or gold or yen or pesos or zorkmids, which may be inflatable by some government, or may be backed only by the full faith and credit of the anonymous remailer in Panama that you reach your ostensibly Cayman-Islands e-bank through. Now for the slight cypherpunks relevance - assuming that you're banking in some currency other than your home country's government's, whether it's hard currency like Swiss Francs or soft metal like gold or a mixture like ECUs or shares of Fidelity Mercury Fund, if you to pay taxes on the net result of transactions, you'll probably want a timestamped log of what you did when, and ideally a good data set of the highest, lowest, and instantaneous prices of the backing currency on several markets, so you can report your profits and losses pessimistically. (For most kinds of accounts, you'd want those sources to be totally separate, but for captive currencies like BankFoo Non-Inflatable Zorkmids, you might want to get them together, though you might want to have two accounts, one of which is quiet and subscribes to the price reports and another where your real transactions happen.) Alice du Gnome-ynous.

1 0

(Fwd) British Study Claims That Photo Credit Cards Don't Work
by Deranged Mutant 17 Dec '03

17 Dec '03

This comes from the FTC Privacy List, and is somewhat relevant.... (The the "Crypto" in the title isn't, oddly...). Some thoughts... (er, questions): 1. What are the implications for log-on systems that rely on recognition of faces (supposedly impossible for hackers to describe and exploit)? 2. Legal implications for witnesses? ------- Forwarded Message Follows ------- Date: Sun, 7 Apr 1996 08:24:18 -0700 From: taxhaven(a)ix.netcom.com (Adam Starchild ) Subject: British Study Claims That Photo Credit Cards Don't Work To: privacy(a)ftc.gov "Crypto-ID" Cards Not Effective Recent studies into the effectiveness of photo credit cards have cast doubt over their ability to cut fraud. Dr. Richard Kemp, of the Department of Psychology at Westminster University, London, organized an experiment involving a London supermarket to test the cards in "the real world." The supermarket was staffed by six people who were all warned to be on the look out for fraudulent credit cards. Dr. Kemp arranged for 44 of his students to pose as shoppers and test the staffs' ability to spot photo-card misuse. Each student was armed with four cards. One showed the student as they were, one showed the student wearing make-up, one showed an individual who vaguely resembled the student and the last card depicted someone who looked nothing like the bearer. People usually recognize photographs of individuals based on a familiarity of the subject. A photograph captures only one angle and expression out of thousands of different combinations. People will recognize photographs of family, friends and well known individuals easily. But how easy is it to accurately compare a photograph with the face of a perfect stranger? At a recent conference, Dr. Kemp said that matching a photo to a stranger's face was "too difficult." He also said that in a non- experimental situation, such as a supermarket, the incidence of fraud detection would be even lower. The results of Dr. Kemp's supermarket experiment proved very interesting. In all, the majority of fraudulent cards were accepted. Amazingly, 35 per cent of the cards bearing a photograph of someone completely different from the student were accepted. A massive 64 per cent of cards bearing a similar individual were also accepted. Another factor which seems to further prove Dr. Kemp's point is that 14 per cent of cards bearing a true likeness of the student were rejected. A few British banks are already offering customers the opportunity to have their photograph etched onto their cards. They claim a reduction in fraud has resulted from this. Dr. Kemp's findings would seem to contradict this belief. Dr. Nicky Towell, one of Dr. Kemp's researchers, said "There is a widely held assumption that photo credit cards are a cheap and effective way of stopping fraud. But this is not the case." No one can tell how well photo credit cards will catch on. But with the majority of people carrying at least one, if not more cards of some sort, how long will it be before photos become compulsory? The government knows that the introduction of ID cards is a political hot potato, but how many people would notice if they turned the cards we already have into crypto-ID cards? Reprinted from The Mouse Monitor, The International Journal of Bureau-Rat Control, a periodical published by Scope International for its customers. Scope International is on the Worldwide Web at http://www.britnet.co.uk/Scope/ Posted by Adam Starchild The Offshore Entrepreneur at http://www.au.com/offshore The privacy list is run automatically by the Majordomo list manager. Send a "help" command to majordomo(a)ftc.gov for assistance. Rob. --- Send a blank message with the subject "send pgp-key" to <WlkngOwl(a)unix.asb.com> for a copy of my PGP key.

2 1

Re: the cost of untracability?
by Frank O. Trotter, III 17 Dec '03

17 Dec '03

> FWIW, I think that there is no capital-gains-type tax on currency > conversions. In other words, if I take dollars and buy yen today, and the > interconvert rate changes and I convert back and make a "profit," that is > not considered income. If that's the case, then ecash has an excellent > precedent behind it to avoid any taxes on interest, especially if that > interest is, in effect, paid by increasing the inherent value of the currency. > > And most of the "interest" will simply be the avoided inflation loss that > would have otherwise occurred. Buying ecash may be equivalent to buying an > absolutely non-inflating currency that the government can't manipulate. > > Jim Bell > jimbell(a)pacifier.com > There is a tax event that occurs when one converts from one currency to another (be it capital or current). In your example the purchase of Yen and later sale may result in a gain if the Yen appreciates, or a loss if it drops. This is a taxable event. You actually refer to the method that was used in the 14th - 16th century in europo to pay interest when it was against church (really "Church") law. Early banks would do currency or metals-to-currency trades with people and imply a rent or interest rate. It was also the begining of the discount trade bill (I'll give you 90 cents today and get your dollar in a year from the person you sold those chairs to). I'll still cling to my Ecash is curency not a currency agument as well BTW. Frank O. Trotter, III - fotiii(a)crl.com www.marktwain.com - Fax: +1 314 569-4906 --------------------------------------------

2 1

Re: CypherPunks -- The Early Years
by nobody＠REPLAY.COM 17 Dec '03

17 Dec '03

Eric Cordian wrote: > > Might I ask why several prominent Cypherpunks have begun reposting > "News of the Weird" through sympatico.ca? Is this some sort of > Pod thing? Close examination of fossil mosquitoes from 40 million years ago show they look just about the same as that one dining on your arm. ============================================== LMBoyd Web Site / U. S. Newspapers / Start Email / Stop Email http://www.LMBoyd.com/postscript.htm

1 0

do it to them. they deseve it.
by bill payne 17 Dec '03

17 Dec '03

Tuesesday March 10, 1998 3/10/98 E-mail Federico.F.Pena(a)hq.doe.gov and mail Federico F. Pena The Secretary of Energy United States Department of Energy Washington, D.C. 20585 Dear Secretary Pena: Purpose of this letter is to appeal a denial of a Freedom of Information Act (FOIA) Request. I wrote on Tuesday February 17, 1999 15:11 e-mail and mail Ms. Elva Barfield Freedom of Information Office U. S. Department of Energy Albuquerque Operations Office/OIEA POB 5400 Albuquerque, NM 87185-5400 EBARFIELD(a)DOEAL.GOV Dear Ms. Barfield: VP Al Gore is in the crypto business. Information SuperSpyWay Al Gore Approved Encryption for China in Return for Campaign Donations by Charles R. Smith Portions of the above document posted on Internet at http://www.us.net/softwar/ and http://www.aci.net/kalliste/ states 1. Gore charged with encryption policy according to PDD-5 and PRD-27 on April 16, 1993. 2. Government officials represent themselves on Al Gore's behalf for RSA patent purchase negotiations in Feb. 1994. 3. RSA chairman Bidzos meets with Chinese officials at the same time as Ron Brown in Oct. 1995. 4. RSA Chairman Bidzos enters into merger negotiations with Security Dynamics, a company backed by Sanford Robertson, in Nov. 1995. 5. VP Gore calls Sanford Robertson from the White House for a donation in Nov. 1995. 6. Robertson delivers $100,000 donation ($80,000 soft - $20,000 directly into the Clinton/Gore campaign) in Jan. 1996. 7. RSA signs deal with China in Feb. 1996. The administration previously prosecuted similar deals but this time does nothing. 8. Justice Dept. approves RSA merger with Security Dynamics in April 1996 for $280 million dollars, netting Sanford Robertson's company a cool $2 million just to write the deal. In 1991 I was in involved with Sandia National Laboratories Comprehensive Test Ban Treaty seismic data authenticator. At that time Sandia director Tommy A Sellers had assumed responsibility for directorship from Robert Clem. Sandia supervisor Tom Wright replaced my supervisor, John Holovka, who was the supervisor for the CTBT seismic data authenticator. Wright brought in Ph.D. Steven Goldsmith to supervise me. Sellars, Wright, and Goldsmith were new to crypto-type projects. Much of this is documented at http://www.jya.com/whp021598.htm. This is evidenced by Sellar's attached SEP 24 1991 memorandum, which Goldsmith help author, addressed to Dr James J Hearn at the National Security Agency. The SEP 24 memorandum contained a number of technical errors. I corrected these errors in my attached December 20, 1991 memorandum. Department of Energy and it predecessors have a well-documented history of not requiring technical expertise for pursuit of interests. Stewart Udall, The Myths of August, writes, Any cover-up must be implemented and enforced by designated agents, and one man emerged in 1953 as the quarterback of the AEC's damage-control effort. His name was Gordon Dunning. Although the personnel charts of the 1950s list him as a low-level "rad-safe" official in the Division of Biology and Medicine, documents demonstrate that he was clothed with authority to manage and suppress information about the radiation released by the testing of nuclear weapons. ... About the time Sellers and Sandia Ombudsman gave me a directed transfer to break electronic locks for the FBI/ERF [engineering research facility], Goldsmith and Wright, certainly with the approval of Sellers, placed a contract with RSA Inc [http://www.rsa.com/] I was told. Ms Barfield, we think the American public needs to know more about RSA's work with Sandia National Laboratories. Therefore, Under the provision of the Freedom of Information Act, 5 USC 552, I am requesting access to: 1 ALL purchase requisitions, including any attached statement of work, issued by Sandia National Laboratories, Los Alamos National Laboratories, or DOE/ALOO between January 1, 1991 and February 17, 1998 to RSA Inc. 2 Copies of all invoices from RSA Inc received by Sandia National Laboratories, Los Alamos National Laboratories, or DOE/ALOO between January 1, 1991 and February 17, 1998 If there are any fees for searching for, or copying, the records I have requested, please inform me before you fill the request. As you know, the Act permits you to reduce or waive the fees when the release of the information is considered as "primarily benefiting the public." I believe that this requests fits that category and I therefore ask that you waive any fees. Your office agreed to waive fees before. This request is surely of "public interest." December 13, 1994 DOE/AL FOIA officer Gwen Schreiner waived fees for the reason, "We have considered your request and have determined that release of the requested records is in the public interest, that disclosure of this information is likely to contribute significantly to public understanding of the operations or activities of the government, that you or the organization you represent have little or no commercial interest in the material contained in the records, that you or the organization you represent have the qualifications and ability to use and disseminate the information, and that the records are not currently in the public domain. A waiver of fees is therefore granted." This waiver of fees was, undoubtedly, issued as a result of former Energy Secretary Hazel O'Leary's Openness initiative. Heart of America paid my way to hear Secretary O'Leary's celebrated whistleblower speech. If all or any part of this request is denied, please cite the specific exemption(s) which you think justifies your refusal to release the information and inform me of your agency's administrative appeal procedures available to me under the law. I would appreciate your handling this request as quickly as possible, and I look forward to hearing from you within 20 working days, as the law stipulates. I received no response to the above FOIA. Therefore I appeal as a result of non-response. As you may be aware (6)(A) Each agency, upon any request for records made under paragraph (1), (2), or (3) of this subsection, shall-- (i) determine within ten days \1\ (excepting Saturdays, Sundays, and legal public holidays) after the receipt of any such request whether to comply with such request and shall immediately notify the person making such request of such determination and the reasons therefor, and of the right of such person to appeal to the head of the agency any adverse determination; and And you may also be aware ------------------------------------------------------------------------- \1\ Under section 12(b) of the Electronic Freedom of Information Act Amendments of 1996 (Pub. L. 104-231; 110 Stat. 3054), the amendment made by section 8(b) of such Act striking ``ten days'' and inserting ``20 days'' shall take effect on October 3, 1997. ------------------------------------------------------------------------- (ii) make a determination with respect to any appeal within twenty days (excepting Saturdays, Sundays, and legal public holidays) after the receipt of such appeal. If on appeal the denial of the request for records is in whole or in part upheld, the agency shall notify the person making such request of the provisions for judicial review of that determination under paragraph (4) of this subsection So I expect a response to this appeal within the time allotted to you by law. Much of this unfortunate matter is appearing on Internet. http://www.jya.com/whp021598.htm http://www.jya.com/cylinked.htm Letter from Bill Payne Regarding Cryptography at Sandia http://www.aci.net/kalliste/ Crypto AG: The NSA's Trojan Whore? http://caq.com/CAQ/caq63/caq63madsen.html NSA, Crypto AG, and the Iraq-Iran War http://www.aci.net/kalliste/speccoll.htm How Secure is America's Nuclear Arsenal? http://www.aci.net/kalliste/nukearse.htm And, of course, we continue to seek settlement. Sincerely, William H. Payne 13015 Calle de Sandias Albuquerque, NM 87111

1 0