From: Heiko Kalte Newsgroups: comp.arch.fpga Subject: Maximum frequency in Virtex and Virtex-E Devices Date: Thu, 04 Jul 2002 09:57:15 +0200 Organization: University of Paderborn, Germany Lines: 25 Message-ID: <3D23FFDB.AB133B84@hni.upb.de> NNTP-Posting-Host: hawaii.uni-paderborn.de Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Trace: news.uni-paderborn.de 1025769433 3890 131.234.86.44 (4 Jul 2002 07:57:13 GMT) X-Complaints-To: usenet@news.uni-paderborn.de NNTP-Posting-Date: 4 Jul 2002 07:57:13 GMT X-Mailer: Mozilla 4.78 [de] (Windows NT 5.0; U) X-Accept-Language: de Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeed.stueberl.de!newsfeed.freenet.de!news.rwth-aachen.de!news.uni-paderborn.de!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:19027 Dear Community, I would like to compare the maximum frequency of the Virtex and Virtex-E family, the databook says 200MHz for the Virtex family and 133MHz (4 LUT levels) for the Virtex-E family. There are obviously different bases for the calculation. Because Virtex-E should be faster. Is there a common basis (e.g. 4 LUT levels) to compare the two series? Heiko -- --------------------------------------------------------------- Dipl. Ing. H. Kalte | HEINZ NIXDORF INSTITUTE | Office: F1.213 System and Circuit Technology | Fon: +49 (0)5251 60-6459 Fürstenallee 11 | Fax: +49 (0)5251 60-6351 33102 Paderborn, Germany | --------------------------------------------------------------- mailto:kalte@hni.uni-paderborn.de http://wwwhni.uni-paderborn.de/sct/ --------------------------------------------------------------- Home of the RAPTOR Rapid Prototyping Systems http://www.RAPTOR2000.de/ --------------------------------------------------------------- ###### From: "enterpoint" Newsgroups: comp.arch.fpga References: <3D23FFDB.AB133B84@hni.upb.de> Subject: Re: Maximum frequency in Virtex and Virtex-E Devices Date: Thu, 4 Jul 2002 15:37:47 +0100 Lines: 51 X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 NNTP-Posting-Host: modem-44-54-60-62.vip.uk.com Message-ID: <3d245cef$1_2@news1.vip.uk.com> X-Trace: 4 Jul 2002 15:34:23 +0100, modem-44-54-60-62.vip.uk.com X-Report: Report abuse to nntpabuse@vip.uk.com Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.icl.net!newsfeed.fjserv.net!newsfeed.icl.net!newsfeed.fjserv.net!iclnet!news1.vip.uk.com Xref: chonsp.franklin.ch comp.arch.fpga:19066 The simple answer to the question is the system performance figure of 240 MHz for the Virtex-E. Beware these figures are only a very rough guide to performance. I have not seen a system figure for Virtex II but significantly faster than both these families is my practical experience. Performance will vary due to a number of factors including coding style of HDL, your synthesiser, the placement table used, routing density, device packing density, pinout and many many more factors. A designer who knows the tools and architecture and uses them effectively can typically achieve up to double the performance over a push button approach of writing HDL and running through the tools. To give an extreme example we have designed a CRC32 generator checker with 3 levels of LUTs which will run at 380 MHz in a Virtex II. This performance however is only obtained by many hours of hand crafting. The same basic design before hand crafting was typically reaching 200 MHz. John Adair Enterpoint Ltd. "Heiko Kalte" wrote in message news:3D23FFDB.AB133B84@hni.upb.de... > Dear Community, > I would like to compare the maximum frequency of the Virtex and Virtex-E > family, the databook says 200MHz for the Virtex family and 133MHz (4 LUT > levels) for the Virtex-E family. There are obviously different bases for > the calculation. Because Virtex-E should be faster. Is there a common > basis (e.g. 4 LUT levels) to compare the two series? > Heiko > > > -- > --------------------------------------------------------------- > Dipl. Ing. H. Kalte | > HEINZ NIXDORF INSTITUTE | Office: F1.213 > System and Circuit Technology | Fon: +49 (0)5251 60-6459 > Fürstenallee 11 | Fax: +49 (0)5251 60-6351 > 33102 Paderborn, Germany | > --------------------------------------------------------------- > mailto:kalte@hni.uni-paderborn.de > http://wwwhni.uni-paderborn.de/sct/ > --------------------------------------------------------------- > > Home of the RAPTOR Rapid Prototyping Systems > http://www.RAPTOR2000.de/ > > --------------------------------------------------------------- ###### Message-ID: <3D246C8E.26ADA2B5@andraka.com> From: Ray Andraka Organization: Andraka Consulting Group, Inc X-Mailer: Mozilla 4.77 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Maximum frequency in Virtex and Virtex-E Devices References: <3D23FFDB.AB133B84@hni.upb.de> <3d245cef$1_2@news1.vip.uk.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Lines: 72 Date: Thu, 04 Jul 2002 15:38:42 GMT NNTP-Posting-Host: 68.15.41.165 X-Complaints-To: abuse@cox.net X-Trace: news1.east.cox.net 1025797122 68.15.41.165 (Thu, 04 Jul 2002 11:38:42 EDT) NNTP-Posting-Date: Thu, 04 Jul 2002 11:38:42 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.ifi.unizh.ch!news.imp.ch!news.imp.ch!newsfeed.stueberl.de!cox.net!p02!news1.east.cox.net.POSTED!53ab2750!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:19038 I think 240 MHz is optimistic for VirtexE, especially considering that the minimum pulse width for the SRL16 is about 238 MHz for a VirtexE-7. It takes an extraordinary designer to get 240 MHz out of VIrtexE in designs containing arithmetic or block memory. Practical top end for VirtexE-6 is around 195 MHz, VirtexE-8 around 240 MHz for very careful design. For 2.5v Virtex, a -4 is about 130 Mhz, a -6 about 160 MHz. Frankly, I've been finding that the VirtexII-4 is about the same performance as a VirtexE-6 as far as arithmetic is concerned for carefully handplaced designs. Stuff that doesn't use the carry chain does appear to be faster in the VirtexII. enterpoint wrote: > The simple answer to the question is the system performance figure of 240 > MHz for the Virtex-E. Beware these figures are only a very rough guide to > performance. I have not seen a system figure for Virtex II but significantly > faster than both these families is my practical experience. > > Performance will vary due to a number of factors including coding style of > HDL, your synthesiser, the placement table used, routing density, device > packing density, pinout and many many more factors. A designer who knows the > tools and architecture and uses them effectively can typically achieve up to > double the performance over a push button approach of writing HDL and > running through the tools. To give an extreme example we have designed a > CRC32 generator checker with 3 levels of LUTs which will run at 380 MHz in a > Virtex II. This performance however is only obtained by many hours of hand > crafting. The same basic design before hand crafting was typically reaching > 200 MHz. > > John Adair > Enterpoint Ltd. > > "Heiko Kalte" wrote in message > news:3D23FFDB.AB133B84@hni.upb.de... > > Dear Community, > > I would like to compare the maximum frequency of the Virtex and Virtex-E > > family, the databook says 200MHz for the Virtex family and 133MHz (4 LUT > > levels) for the Virtex-E family. There are obviously different bases for > > the calculation. Because Virtex-E should be faster. Is there a common > > basis (e.g. 4 LUT levels) to compare the two series? > > Heiko > > > > > > -- > > --------------------------------------------------------------- > > Dipl. Ing. H. Kalte | > > HEINZ NIXDORF INSTITUTE | Office: F1.213 > > System and Circuit Technology | Fon: +49 (0)5251 60-6459 > > Fürstenallee 11 | Fax: +49 (0)5251 60-6351 > > 33102 Paderborn, Germany | > > --------------------------------------------------------------- > > mailto:kalte@hni.uni-paderborn.de > > http://wwwhni.uni-paderborn.de/sct/ > > --------------------------------------------------------------- > > > > Home of the RAPTOR Rapid Prototyping Systems > > http://www.RAPTOR2000.de/ > > > > --------------------------------------------------------------- -- --Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email ray@andraka.com http://www.andraka.com "They that give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -Benjamin Franklin, 1759 ###### Message-ID: <3d25ac8d$0$14695$afc38c87@news.optusnet.com.au> From: hamish@cloud.net.au Subject: Re: Maximum frequency in Virtex and Virtex-E Devices Newsgroups: comp.arch.fpga References: <3D23FFDB.AB133B84@hni.upb.de> <3d245cef$1_2@news1.vip.uk.com> <3D246C8E.26ADA2B5@andraka.com> User-Agent: tin/1.5.13-20020614 ("Sugar") (UNIX) (Linux/2.2.18 (i586)) Date: 05 Jul 2002 14:26:21 GMT Lines: 15 NNTP-Posting-Host: 211.28.158.161 X-Trace: 1025879181 14695 211.28.158.161 Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!news1.optus.net.au!optus!spool01.syd.optusnet.com.au!spool.optusnet.com.au!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:19063 Ray Andraka wrote: > Frankly, I've been finding that the VirtexII-4 is about the same performance as > a VirtexE-6 as far as arithmetic is concerned for carefully handplaced designs. > Stuff that doesn't use the carry chain does appear to be faster in the VirtexII. I don't do much arithmetic, so I've found the VirtexII much faster than the E. I have noticed that carry chain entry/exit times are much higher than the VirtexE equivalents, making carry chains relatively expensive for small functions. As a result, a couple of carry chains in series between flip flops worked okay at 160MHz in Virtex-E but not at the same speed in Virtex-II. Hamish -- Hamish Moffatt VK3SB ###### Message-ID: <3D25BBA4.AFC2132D@andraka.com> From: Ray Andraka Organization: Andraka Consulting Group, Inc X-Mailer: Mozilla 4.77 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 Newsgroups: comp.arch.fpga Subject: Re: Maximum frequency in Virtex and Virtex-E Devices References: <3D23FFDB.AB133B84@hni.upb.de> <3d245cef$1_2@news1.vip.uk.com> <3D246C8E.26ADA2B5@andraka.com> <3d25ac8d$0$14695$afc38c87@news.optusnet.com.au> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 40 Date: Fri, 05 Jul 2002 15:28:22 GMT NNTP-Posting-Host: 68.15.41.165 X-Complaints-To: abuse@cox.net X-Trace: news1.east.cox.net 1025882902 68.15.41.165 (Fri, 05 Jul 2002 11:28:22 EDT) NNTP-Posting-Date: Fri, 05 Jul 2002 11:28:22 EDT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeed.stueberl.de!cox.net!p02!news1.east.cox.net.POSTED!53ab2750!not-for-mail Xref: chonsp.franklin.ch comp.arch.fpga:19039 We do mostly arithmetic since we specialize in DSP applications of FPGAs. Between the slow multiplier speeds in V2 before those produced this spring and the dreadfully slow times getting on and off the carry chains, and the smaller number of slices per marketing gate (and indirectly per dollar) we have found the VirtexE to be generally a much better value for DSP applications. Even with the fixed multiplier speeds, you will presumably be using other arithmetic in conjunction with the multipliers, and at 18 bits your speed gets limited by the carry chain to something about equivalent to VirtexE. hamish@cloud.net.au wrote: > Ray Andraka wrote: > > Frankly, I've been finding that the VirtexII-4 is about the same performance as > > a VirtexE-6 as far as arithmetic is concerned for carefully handplaced designs. > > Stuff that doesn't use the carry chain does appear to be faster in the VirtexII. > > I don't do much arithmetic, so I've found the VirtexII much faster than > the E. I have noticed that carry chain entry/exit times are much higher > than the VirtexE equivalents, making carry chains relatively expensive > for small functions. As a result, a couple of carry chains in series between > flip flops worked okay at 160MHz in Virtex-E but not at the same speed > in Virtex-II. > > Hamish > -- > Hamish Moffatt VK3SB -- --Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email ray@andraka.com http://www.andraka.com "They that give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." -Benjamin Franklin, 1759