To: Khris Vogel who wrote (52876 ) 4/14/1998 2:45:00 AM From: Paul Engel Read Replies (3) | Respond to of 186894
Intel Investors - First Benchmark Reports of 100 Mhz 440BX/400 MHz Pentium II devices. Check this out. They note a 21% increase in performance, using BAPCO benchmarks, in going from a 333 MHz PEntium II (66 MHz Bus Speed) to a 400 MHz Pentium II (100 MHz Bus speed). Note that (400-333)/333 = 67/333 = 20%. Thus, the 20% "expected" speed improvement in going from 333 to 400 Mhz was indeed realized - since now the external bus and memory system were also speeded up by use of the 100 MHz 440BX and SDRAM memory. Paul {=================}techweb.com copyright © CMP Proteva 400MHz Is a Screamer By Kristen Kenedy San Mateo, Calif. - When PCs with 350MHz and 400MHz CPUs begin shipping later this week, retailers will have more to sell than faster processors. Along with the step-up in chip speed, Intel is adding a 100MHz system bus, up from 66MHz in systems with 333MHz and slower CPUs. That's close to a 50 percent increase in the rate data moves among the processor, its main memory and various components. Intel maintains that PC users will see a performance improvement of greater than 20 percent with the new architecture. Bapco SysMark tests used at CRW Labs reported an average increase of about 21 percent using a 400MHz system over a similarly configured 333MHz PC. CPU speed and bus aren't the only determining factors of system performance, yet a comparison of 333MHz and 400MHz systems gives the market an early indication of overall performance expectations. We took a look at a prototype Proteva-branded system with an Intel Pentium II 400MHz CPU, 32MB of RAM, a 6.4GB IDE Western Digital hard drive, and an early version of an AGP graphics card based on the Intel 740i graphics chip with 4MB of RAM on board. We compared it to existing tests we ran on a PC with a 333MHz CPU, 32MB of RAM, a Quantum 6.4GB hard drive and an AGP ATI Rage Pro with 4MB of RAM on board. We found performance benefits of the 400MHz chip were readily apparent. Graphic-intensive applications, such as Adobe PageMaker and CorelDraw, performed about 26 percent and 24 percent better than on a 333MHz PC, respectively. The advantage of having a 100MHz system bus was also quickly apparent. Information that comes from the PCI bus (where devices such as sound cards reside), hard drives and optical drives now travel toward the CPU and main memory at almost twice the speed. The new processors also bring an increase in transfer speeds on the Accelerated Graphics Port. Where many systems were running at 1x speeds, or 66MHz, new systems with the Intel 440BX chip set can move data to and from the AGP bus at 133MHz. In theory, that means a leg-up for graphics-intensive applications, such as 3-D games. Flight simulators, for example, that feature moving clouds and varying terrains require some initial set-up processing in the CPU and then move to the graphics card for rendering. With a faster system and AGP bus, the data transfer speeds should increase exponentially. Nonetheless, we saw only a slight performance increase in our 3-D test game, Quake II. The reason: Quake II wasn't designed for AGP, so it doesn't take advantage of some of the performance enhancements. The system was able to produce 19.3 frames per second during our test, higher than average 2-D/3-D AGP cards we've tested, but not better than the combination of separate 2-D and 3-D PCI accelerator boards we've seen in the past. However, we predict that games written specifically for the platform will fly on any new 400MHz system. For immediate gratification, users should check out soft-DVD playback on a 400MHz system. We detected slight frame skipping on an early 400MHz demo, but colors were vibrant and playback was good enough for general consumption. Other consumer applications, such as image editing, will also receive a performance boost. We recommend using the usual qualification process when selling high-end systems at retail. Pitch 400MHz to customers who want top-quality goods or who seek "future-proof" systems that can handle applications that haven't even been introduced yet. Methodology CRW used Bapco's SysMark 32 benchmark to determine performance on a number of industry-standard applications. We also ran a frame count on a Quake demo, demo1.dm2, to look at multimedia functionality. All tests were run three times and then averaged for a final number. Copyright (c) 1998 CMP Media Inc. New Search | Search the Web