X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f From: "Rod Pemberton" Newsgroups: comp.os.msdos.djgpp Subject: Re: Performance enhancement for gettimeofday()? Date: Sat, 13 Jan 2007 18:18:48 -0500 Organization: Aioe.org NNTP Server Lines: 89 Message-ID: References: <5t3iq2dlv4hoh2sdf9c4dbot50e6ge28mn AT 4ax DOT com> NNTP-Posting-Host: IVw7K97ih4IohxRqyKkqFw.user.aioe.org X-Complaints-To: abuse AT aioe DOT org X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1441 X-Priority: 3 X-Newsreader: Microsoft Outlook Express 6.00.2800.1437 X-MSMail-Priority: Normal To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Reply-To: djgpp AT delorie DOT com "Brian Inglis" wrote in message news:5t3iq2dlv4hoh2sdf9c4dbot50e6ge28mn AT 4ax DOT com... > On Wed, 10 Jan 2007 13:25:42 -0500 in comp.os.msdos.djgpp, "Rod > Pemberton" wrote: > > > > > wrote in message > >news:OF77475379 DOT 7BA371D5-ON8725725F DOT 00598107-8725725F DOT 0059B5ED AT seagate DOT com. .. > >> Rod Pemberton wrote on Tue, 9 Jan 2007 at 03:46:03 -0500: > >> > >> # I believe this it the math you'll need: > >> # > >> # 14.318Mhz=4*3.58Mhz=4*(4.5Mhz*455/572) > >> # (4.5Mhz US TV bandwith/channel, 455 colorburst phase changes/line, > >> 572 > >> # total lines/frame including sync) > >> # 14.318Mhz/12=1.93182Mhz > >> > >> Aha, this is the one that's why our numbers don't agree: > >> 14.318MHz divided by 12 is actually 1.193666... MHz. > >> > > > >Sorry, it appears I failed to type a 1 following the decimal. It's not > >14.318000MHz, but 14.318181MHz. You really need to enter > >4*4.5*(10^6)*455/572 to compute the 14.318MHz and work from there. IIRC > >('twas 25+ years ago), it's 4 times the colorburst as calculated by the > >original engineer who designed the US color TV standard. That way you won't > >loose precision. Of course, a real crystal usually has a tolerance range, > >but that range is usually small compared to the frequency, like +/- 100Hz or > >+/-10KHz. Of course, you could go to Mouser or another electronic supplier, > >and look for a crystal if you think the range would help. > > > >Like you, I'll use ... for repeating digits. The 1 and 8 repeat for both. > >I was using more decimals but rounded/truncated. > > > >14.318181818181... Mhz / 12 = 1.193181818181... Mhz. > >1.193181818181...Mhz / 65536 = 18.206509676846 Hz > > IIRC crystal frequency 157.5MHz = 9/2*7*5*1E6, /11 colour burst A 157.5Mhz crystal in 1980's? ROFL! In a PC, in the 1980's? ROFL! (Where's the Kleenex, I've got to wipe away the tears...) It's possible that 157.5Mhz crystals are being used in PC's today. It's plausible, but I doubt it (I haven't looked lately.). The primary factor is cost. For PC's, the lowest cost solution usually wins. I know of only three PC exceptions to that rule. A 157.5Mhz crystal costs more than a 14.318Mhz crystal, and only provides an advantage _if_ the PC uses needs an other clock closely related to 157.5Mhz. With common PC bus frequencies of 66, 100, 133, 166, 200, 215, and 225Mhz, 157.5Mhz _seems_ to be a really _odd_ frequency choice to me. I had schematics for a few '80's PC's during the '80's: C64, Apple II, etc. The highest frequency in the schematic was always some multiple of the colorburst which was used by the video circuitry. If you actually have a real schematic showing a that 157.5Mhz crystal was in use in a 1980's era PC of any common make (i.e., no Crays), it'd probably be worth a fortune... Unfortunately, I wasn't actively monitoring the advancement in crystal oscillator frequencies. However, if you're interested, this link contains an oscillator frequency timeline: http://www.npcamerica.com/Datasheets/KEYNOTE2.PDF > 14318181.8..Hz, /3 for PC clock 4772727.27..Hz, /4 for timer > 1193181.8..Hz, /65536 for tick 18.2065096768465909..Hz, giving a period > of 54925.4095238..us with the last six digits repeating. > > So the relevant factors here are 7*5*3/11/2^19. > Sure, except if one of those factors came from the 157.5Mhz. As I showed, the colorburst and the common 4*colorburst crystal are derived from the TV baseband, created for the US color TV standard to be compatible with US Black & White (BW) TV's, i.e., not from 157.5Mhz. Rod Pemberton