Shareware Grab Bag

home *** CD-ROM | disk | FTP | other *** search

/ Shareware Grab Bag / Shareware Grab Bag.iso / 090 / v20jr.arc / V20JR.TXT < prev next >

Wrap

Text File | 1986-03-08 | 18.1 KB | 381 lines

PCjr V20 SPEEDUP The IBM PCjr's Performance With the Intel 8088 vs. the NEC V20 How fast is the new NEC V20 replacement microprocessor chip? This chip imitates the Intel 8088 and is virtually 100% compatible,but operates more efficiently. I recently replaced the Intel 8088 chip in my PCjr with a V20. My PCjr has the Racore "Drive Two" second disk drive, 512K internal random-access memory (RAM), and the Racore direct memory access (DMA) board. The V20 can also be installed successfully on a stock PCjr. To find out how much, if any, the V20 improved my system's performance, I conducted a series of tests. I tested five configurations: A stock PCjr (using the PCjr switch setting on the Racore unit), with both an 8088 and a V20; the same PCjr switched to Racore's PC mode, using DMA, with both an 8088 and a V20; and an IBM PC/XT with 640K and an 8088, using its floppy disk drive for the disk input/output tests. The tests also demonstrated how much the Racore Drive Two and DMA system improves the PCjr's performance and how the speed of the four PCjr configurations compares to an IBM XT using its floppy diskette drive. The Benchmark Tests First I used the "System Information" program in the Norton Utilities Version 3.0, which gives a machine's "computing performance index relative to IBM/PC." A stock PC gets an index of 1.0 in this test. Then I ran four of the benchmark programs PC Magazine uses to test computers. (See a description of these PC programs in Note 1 at the end of this file.) Conclusions The NEC V20 delivers a reasonable performance improvement for a small price ($16 to $25 by mail). But a Racore Drive Two- based system with additional memory delivers a more significant performance gain over a stock PCjr, even without a V20; of course, you pay a $500 premium for that package. If you want to speed up your PCjr without spending much, replace its 8088 with a V20. If you want to make your PCjr as fast as a PC, add the Racore expansion. If you want a PCjr that's even faster than a PC, add the V20 to your Racore/PCjr system. The NEC V20's most stunning showing came in the Norton Utilities performance index. An stock PCjr performed 28.6% better with the V20 than with the 8088. But the expanded PCjr, with the Racore "SPEEDMA" speedup program and DMA, was 80% faster with the V20 than without it, and 80% faster than the PC/XT with an 8088! Page 2 PCjr V20 SPEEDUP October 17, 1985 The stock PCjr showed a 10.6% improvement and the Racore- expanded PCjr a 20.3% improvement running the BASIC benchmarks with the V20 chip. The disk version of Lotus 1-2-3 ran 10.7% faster with the V20 on the expanded PCjr. The diskette operation tests produced mixed results. In one test the V20 increased the PCjr's speed by as much as 33.7%, but in another set of tests the PCjr was as much as 12.5% slower with the V20. I ran these tests several times to confirm the results. (See Note 2.) I have no explanation for these seemingly contradictory results. The stock and expanded PCjrs were 17.6% and 33.7% faster, respectively, with the V20 chip in PC Magazine's BASIC file update benchmark. In fact, the expanded V20-equipped PCjr was 12.4% faster than the XT in this test. But the fastest combined results in PC's disk input/output benchmarks came on the expanded PCjr WITHOUT the V20 chip; it was faster than the XT. The V20 seemed to slow down the expanded jr in these tests. The stock jr was marginally faster with the V20 than with the 8088. The V20 improved the PCjr's performance in the prime numbers calculations by less than 6%. The winner here was the XT. Finally, the PCjr with the V20 was the Lotus 1-2-3 speed champion, delivering a 10.7% improvement over the 8088 PCjr. The surprise here was a 12.8% advantage over the XT running the same version of 1-2-3. In fact, even the expanded PCjr with an 8088 was a bit faster than the XT in this test. Here is a summary of the performance changes I observed after adding the V20: Adding V20 to Adding V20 to 128K PCjr 512K PCjr with without Racore Racore & DMA Norton's Performance Index 28.6% 80.0% Disk I/O Benchmarks 0.3% -12.5% Prime Number Benchmark 5.9% 5.6% BASIC RAM-based Benchmarks 4.1% 5.2% BASIC File update Benchmark 17.6% 33.7% Lotus 1-2-3 Benchmark NA 10.7% Page 3 PCjr V20 SPEEDUP October 17, 1985 The Champion Of the five computer configurations I tested, the expanded V20 PCjr (with Racore's SPEEDMA program and DMA) was the winner. To summarize the results, I have assigned five points for each first place, four for each second, and so on. It took first place in four of the six types of benchmarks and scored 27 out of a possible 30 points. The expanded 8088 PCjr and the XT tied with 22 points and each took one first place (of course, the XT would do much better with in the disk tests with its hard disk). Here is a summary of the rankings in the five sets of tests: PCjr w/ PCjr w/ PCjr w/ PCjr w/ Intel NEC Intel NEC PC/XT 8088 & V20 & 8088 & V20 & w/Intel Racore Racore Racore Racore 8088 w/ Drive 2, Drive 2, Drive 2 Drive 2 Floppy No DMA, No DMA, w/ DMA, w/ DMA, Disk jr Mode jr Mode PC Mode PC Mode & DMA ----------------- ----------------- -------- Norton's Performance Index 5 4 2 1 2 Disk I/O Benchmarks 5 4 1 3 2 Prime Numbers Benchmark 5 4 3 2 1 BASIC RAM 5 4 2 1 3 BASIC File 5 3 4 1 2 Lotus 1-2-3 Worksheet 2 1 3 Total Scores (30 possible) 5 11 22 27 22 Here are the results of the benchmarks. First I'll list the raw data from the tests of the five machine configurations. Then I'll compare some of the results: PCjr w/ PCjr w/ PCjr w/ PCjr w/ Intel NEC Intel NEC PC/XT 8088 & V20 & 8088 & V20 & w/Intel Racore Racore Racore Racore 8088 w/ Drive 2, Drive 2, Drive 2 Drive 2 Floppy No DMA, No DMA, w/ DMA, w/ DMA, Disk jr Mode jr Mode PC Mode PC Mode & DMA ----------------- ----------------- -------- NORTON UTILITIES 3.0 Computing Performance Index Relative to IBM PC 0.7 0.9 1.0 1.8 1.0 DISK I/O BENCHMARKS Random Write 128K 67.1 55.1 51.8 61.0 61.7 300K 73.5 77.5 63.7 77.9 42.1 512K 53.1 55.4 39.9 54.5 63.2 Page 4 PCjr V20 SPEEDUP October 17, 1985 578K 61.0 59.3 48.7 54.8 63.2 ------- ------- ------- ------- ------- 254.7 247.4 204.0 248.3 230.2 Random Read 128K 56.5 57.6 49.8 49.9 46.4 300K 106.8 104.4 91.3 91.4 85.0 512K 68.3 68.2 59.9 63.1 51.1 578K 136.9 138.4 123.4 122.5 116.3 ------- ------- ------- ------- ------- 368.5 368.6 324.4 326.9 298.9 Sequential Read 128K 13.6 14.2 7.5 9.7 9.0 300K 30.2 30.8 22.4 31.1 15.8 512K 34.4 36.7 24.4 34.0 38.8 578K 44.1 45.3 31.5 41.2 46.4 ------- ------- ------- ------- ------- 122.3 127.0 85.7 116.0 110.0 Summary of Disk I/O Benchmarks: Random Write 254.7 247.4 204.0 248.3 230.2 Random Read 368.5 368.6 324.4 326.9 298.9 Sequential Read 122.3 127.0 85.7 116.0 110.0 ------- ------- ------- ------- ------- 745.4 743.0 614.2 691.2 639.0 PRIME NUMBERS 1-50 51 48 45 43 35 BASIC BENCHMARKS Integer Addition #1 65 62 50 47 52 Integer Addition #2 52 50 40 38 42 Floating Point 150 143 115 108 120 String Concatenation 114 110 90 86 95 Data Lookup 162 156 125 119 131 Empty Loop 13 12 10 9 10 ------- ------- ------- ------- ------- BASIC RAM-based benchmarks 556 533 430 407 450 BASIC File Update 505 416 481 319 364 ------- ------- ------- ------- ------- BASIC Benchmarks, Total 1,061 949 911 726 814 LOTUS 123 Worksheet 244 218 250 Page 5 PCjr V20 SPEEDUP October 17, 1985 PCjr w/ PCjr w/ Intel NEC 8088 & V20 & Racore Racore Percent Drive 2, Drive 2, Speed No DMA, No DMA, Increase/ jr Mode jr Mode (Decrease) ----------------- ---------- NORTON UTILITIES 3.0 Computing Performance Index Relative to IBM PC 0.7 0.9 28.6% DISK I/O BENCHMARKS 745.4 743.0 0.3% PRIME #'s 1-50 51 48 5.9% BASIC RAM-based benchmarks 556 533 4.1% BASIC File Update 505 416 17.6% ------- ------- -------- BASIC Benchmarks, Total 1,061 949 10.6% LOTUS 123 Worksheet PCjr w/ PCjr w/ Intel NEC 8088 & V20 & Racore Racore Percent Drive 2 Drive 2 Speed w/ DMA, w/ DMA, Increase/ PC Mode PC Mode (Decrease) ----------------- ---------- NORTON UTILITIES 3.0 Computing Performance Index Relative to IBM PC 1.0 1.8 80.0% DISK I/O BENCHMARKS 614.2 691.2 -12.5% PRIME #'s 1-50 45.0 42.5 5.6% BASIC RAM-based benchmarks 430 407 5.2% BASIC File Update 481 319 33.7% ------- ------- ------- BASIC Benchmarks, Total 911 726 20.3% LOTUS 123 Worksheet 244 218 10.7% Page 6 PCjr V20 SPEEDUP October 17, 1985 PCjr w/ PC/XT Intel w/Intel 8088 & 8088 w/ Racore Percent Floppy Drive 2 Speed Disk w/ DMA, Increase/ & DMA PC Mode (Decrease) ----------------- --------- NORTON UTILITIES 3.0 Computing Performance Index Relative to IBM PC 1.0 1.0 0.0% DISK I/O BENCHMARKS 639.0 614.2 3.9% PRIME #'s 1-50 35.0 45.0 -28.6% BASIC RAM-based benchmarks 450 430 4.6% BASIC File Update 364 481 -32.1% ------- ------- ------- BASIC Benchmarks, Total 814 911 -11.9% LOTUS 123 Worksheet 250 244 2.4% PCjr w/ PC/XT NEC w/Intel V20 & 8088 w/ Racore Percent Floppy Drive 2 Speed Disk w/ DMA, Increase/ & DMA PC Mode (Decrease) ----------------- ---------- NORTON UTILITIES 3.0 Computing Performance Index Relative to IBM PC 1.0 1.8 80.0% DISK I/O BENCHMARKS 639.0 691.2 -8.2% PRIME #'s 1-50 35.0 42.5 -21.4% BASIC RAM-based benchmarks 450 407 9.6% BASIC File Update 364 319 12.4% ------- ------- ------- BASIC Benchmarks, Total 814 726 10.8% LOTUS 123 Worksheet 250 218 12.8% Page 7 PCjr V20 SPEEDUP October 17, 1985 Note 1: PC Magazine uses more than a dozen standard benchmark programs its staff has developed to test and compare personal computers. I recently downloaded a library of those programs (either from CompuServe or the Exec-PC BBS) and used four of them in my tests. Here is a brief description of the benchmarks: a. BENCH01.EXE, "Disk I/O," which performs random-write, random-read and sequential-read disk input/output tests using a 200K file comprised of 128K, 300K, 512K and 578K records; b. BENCH02.EXE, "Prime Number Calculation;" c. BENCH12.BAS, "Quick BASIC Programs," which test a computer's speed running RAM-based BASIC integer addition, floating point, string concatenation, data lookup, and empty loop programs, and a disk I/O-based BASIC file update program; and d. BENCH13.WKS, "Application Testing - Lotus 1-2-3 Spreadsheet," in which a user times how fast a computer can run several macro-driven operations in a Lotus 1-2-3 spreadsheet. Note 2: I'm not a technician or computer expert, and I don't represent these benchmarks to be 100% accurate. I attempted to ensure consistency and, in most cases, ran each test two or three times and averaged the results, but there may have been subtle differences of which I am unaware in the conditions under which I ran the tests before and after installing the NEC chip. So use this information as a guide, but don't consider it to be definitive. In most cases, the performance differences are substantial enough that I suppose they would be statistically significant even if a margin of error were assumed.