G
Guest
Guest
GOBBLE,
WM7 Does NOT show the 'TRUE' performance of the AthlonXP in the real world. But rather the performance of the chip in the ONLY APP ON THE FREAKIN PLANET that gets it all wrong. Unless you have other apps you can name in the real world that exhibit this behavior, you should bite your tongue. Here's my evidence for the REAL WORLD. Don't forget that TOM uses INTEL's RECOMPILED code for SpecViewPerf. Which is NOT REAL WORLD. Those apps are NOT recompiled themselves, only the benchmark shows the performance improvement. The exact OPPOSITE of the AMD situation. Where in the real world the apps DO SEE THE SSE ON AthlonXP. ONLY the app in sysmark 2000/2001 has a problem. Also don't forget BAPCO/MadOnion is owned by Intel. Despite whatever they say. INTEL originally registered/owned their DOMAIN NAME, they live on Intel land (free rent?, favored benchmarks anyone?..LOL)...once word got out Intel dumped any association with them, yet bias still shows in the benchmark (65% weighting for WM7 for one APP? that happens to screw AMD? used in webmark2001 and sysmark2001). Anyway here's my previous post at HARDOCP.
FOLLOW THIS DATA PEOPLE:http://www.anandtech.com/cpu/showdoc.html?i=1543&p=6
QUOTE: Enabling SSE gives the Athlon XP an incredible 18% increase in performance. This is a real world performance increase courtesy of SSE. [ internet content creation portion of Sysmark 2001]
Overall Sysmark from same Anandtech.com page
QUOTE:"The overall performance of the Athlon XP is improved by a little over 8% because we enabled SSE.
You can clearly see that enabling SSE paints a much brighter picture for the Athlon XP; arguably a much truer picture of its performance."
[did you guys get that much truer picture part? Note ONLY WME7(the problem app) can't see SSE unless patched, the rest of the SSE capable apps in here already id the SSE in athlonXP just fine if they are capable of SSE]
Note Intel (of course, and Intel's right hand Bapco) are ticked about this. Though Intel had no problem competely changing/recompiling code for SSE2 in SpecViewPerf.
Same Anandtech article few pages later about SpecViewPerf 6.1.2 DX-6 performance:
QUOTE:
"The Pentium 4 is much more competitive in this viewset but more importantly is the huge performance increase that is courtesy of the Palomino core enhancements. You can see that the Athlon 1.4GHz is clearly the worst performer in this benchmark, but the Palomino core enhancements give the Athlon XP 1.4GHz processor a 34% advantage over its older brother." [ While a bit (likely 2-10%, if old comments on AthlonMP in previous articles are correct) is coming from the TLB enhancements, most of the 34% is coming from SSE. Remember the comparison is same speed athlon vs athlon XP: You also see in benchmarks that don't use SSE athlon XP at same speed as Tbird 1.4 only nets a scant 2-10% or so...core enhancements other than SSE aren't worth the huge gains seen in some benchmarks ]
[ Nothing to do with SSE here, but I couldn't resist throwing in a quote from Techreport.com about "the internet is faster on P4" crap Kyle chants (webmark is awesom...HONEST...LOL) ]
QUOTE:Before the Intel fanboys start hooting, they should pause to consider Intel's own use of loopy marketing terms. NetBurst, anyone? That one is chosen to correspond with Intel's whole "makes the Internet go faster" line, which is a load of bunk. [That can be found here: http://www.tech-report.com/reviews/2001q4/athlonxp/index.x?pg=4 ]
TechReport.coms Duron 1ghz (morgan with SSE) review, check out the SpecViewPerf scores of a 1ghz Duron regularly beating a larger cached Tbird 1ghz: SSE anyone (we know this benchmark is optimized for both SSE and SSE2?
http://www.tech-report.com/reviews/2001q3/duron-1ghz/index.x?pg=7
Check out AcesHardware.com's piece about video2000 and SSE, when a MORGAN 1ghz (38.1) tramples a Tbird 1ghz (29.8) when encoding mpeg2 videos. Link: http://www.aceshardware.com/Spades/read.php?article_id=45000210
Here's the QUOTE: "PCT reports that about 15% of of the instruction mix consists of MMX instructions. However, SSE-instructions are also reported as "MMX." If you take a look at a previous MPEG2-encoding benchmark , you will notice that a Pentium III 800 performs like an Athlon 1000. This indicates that many of those MMX instructions are, in fact, SSE instructions. The Morgan on the MSI KT266 Pro performed much better than on the ABIT KT7A because of the higher memory bandwidth (DDR), but more importantly, because the DVD2000 benchmark is SSE optimized." [More from same page] "Keeping Morgan's MPEG-2 performance in mind, we can easily predict that the upcoming desktop Palomino will take the top spot in such SSE (non SSE-2) optimized applications." [ LOOK Ma- No patch required for SSE! But, I'll find some way to turn it off...I don't want all that FREE PERFORMANCE]
Photoshop at VansHardware.com (Intel Benchmarketing 101 article : bapco includes HEAVY SSE, but EXCLUDES ANY 3DNOW): Note that nothing has to be done to get SSE in photoshop either! Uh, real world performance at work AGAIN.
QUOTE:"Despite its 200 MHz handicap, the 1.2 GHz Palomino easily defeated its faster clocked older brother. Although the Palomino also supports hardware prefetching, the performance difference is almost entirely due to SSE since the SysMark Photoshop test is relatively bandwidth insensitive as can be seen in this old chipset review.
Our results demonstrate that the 1.2 GHz Palomino has nearly a 30% better score than the 1.4 GHz Thunderbird in the SysMark 2000 Photoshop test.
Last year, this SSE advantage would belong solely to Intel's Pentium III. VansHardware.com LINK:
http://www.vanshardware.com/articles/2001/august/010814_Intel_SysMark/010814_Intel_SysMark.htm"
[SSE allows 200mhz slower AthlonMP 1.2 to SMOKE a Tbird 1.4!...did I mention you don't need a patch? It works in the real world ]
[Extremetech.com AthlonXP article (Loyd Case) MpegPro2 scores cut in half by 1.53XP over 1.4 Tbird! ] Quote: "MPPRO2 has been enhanced for SSE, and it seems that the Athlon XP with its new SSE capability, takes good advantage of those optimizations"
[ Well, it couldn't be that SSE crap working in the REAL WORLD again could it? Oh, and without a patch for gods sake! NO WAY!] LINK:http://www.extremetech.com/article/0,3396,apn=11&s=1005&a=16008&app=9&ap=10,00.asp
[ Same article at extremetech.com about MusicMatch 6.1 with fullblown SSE! ]
QUOTE: "Again, the Athlon XP scores big over the Pentium 4, taking advantage of the SSE enhancements built into MusicMatch 6.1, cutting the P4's time by nearly half."
[ The 1.53XP scores well over 100% faster than the Tbird 1.4! I know, this SSE stuff doesn't work in the REAL WORLD right Kyle? Did I forget to mention it NEEDS NO PATCH? ]
Since I'm tired of this and I FULLY think my point was proven, I'll quit collecting data now. All we need to know now is why Kyle insists on attempting to DUPE his readers? I'm totally open to KYLE's RESPONSE HERE. No I don't want your Intel BS. Give me something that you can PROVE. Like SSE not working in a patched WME7 after your done benchmarking...LOL. Or a list of apps that won't work with AthlonXP's SSE. Do anything, but PROVE SOMETHING. I only see you telling people 'i stand by my bs, even if I can't prove any of it'. KYLE: "Giving people the wrong expectations of current software is simply wrong and I will stand my ground on this one." ....[I guess you better look beneath your feet KYLE. Your WALKING ON AIR pal. In order to "stand you ground" you actually have to have some GROUND to stand on don't you? ]
JOJIT: Is this enough data for you? I don't believe I have to KNOW a person [personally] to know that the data the person is putting forth is an outright LIE. Wait a minute. He had NO DATA. Also, since ID software never released the AMD K62 patch for Quake2 (AMD DID IT) why did everyone use it in benchmarks after it was released? Is it NOT enough that microsoft helped fix it? I actually believe this was a condition microsoft/amd agreed upon. Something to help SELL WinXP and AthlonXP together. AMD is marketing AthlonXP as the best thing since sliced bread with WinXP. Microsoft will make the patch incorporated into WinXP's WME (probably already did). Just a guess here.
However, NONE of my data above is GUESSING. The numbers DON'T LIE. SSE is used in the REAL WORLD on AthlonXP. As such, using the patch in WME7 is giving an ACCURATE representation of performance when using AthlonXP. I might think differently if it was some kind of BETA thing that wouldn't run on some PC's, or it could crash Win2K causing you to reformat (visions of Roxio Easy CD Platinum wiping out hard drives everywhere here..ROFL). However this is JUST an SSE switch for that which is ALREADY built into the software. Has NO changes to the code. It will crash NO PC. Harms nothing.
Any proof from KYLE would be great. I'd love to take back all I said. I really liked this site. Right now I'm rethinking this, as should all of you! I will retract all LIE words if Kyle can prove to us his RESPONSE to ENKAFAN was accurate.
ENKAFAN asked if the patch ONLY improves the benchmark. To which KYLE replied: Basically YES, at this time.
Blatant LIE. Which I proved MORE than adequately above. In order for him to be even remotely right (even just regarding Windows Media Encoder), SSE would have to die on Athlon as soon as you quit benchmarking. This is again, UNTRUE. The patch will work forever after you do it. Also it ONLY applies to the one APP [WME7] that happens to SKEW 2 benchmarks (webmark 2001, and Sysmark 2001) totally in Intel's favor. The rest of the REAL WORLD, ID's a CPU's SSE BIT, and NOT (like WME7) as the CPUID string. As they say: Proof is in the pudding. Wheres your proof KYLE?
Shibumi
PS:KYLE BANNED ME AFTER THIS BUT DIDN'T RESPOND TO THE DATA. Even said I could get sued for this...ROFL. For telling the truth? Is that a crime today? I digress.
WM7 Does NOT show the 'TRUE' performance of the AthlonXP in the real world. But rather the performance of the chip in the ONLY APP ON THE FREAKIN PLANET that gets it all wrong. Unless you have other apps you can name in the real world that exhibit this behavior, you should bite your tongue. Here's my evidence for the REAL WORLD. Don't forget that TOM uses INTEL's RECOMPILED code for SpecViewPerf. Which is NOT REAL WORLD. Those apps are NOT recompiled themselves, only the benchmark shows the performance improvement. The exact OPPOSITE of the AMD situation. Where in the real world the apps DO SEE THE SSE ON AthlonXP. ONLY the app in sysmark 2000/2001 has a problem. Also don't forget BAPCO/MadOnion is owned by Intel. Despite whatever they say. INTEL originally registered/owned their DOMAIN NAME, they live on Intel land (free rent?, favored benchmarks anyone?..LOL)...once word got out Intel dumped any association with them, yet bias still shows in the benchmark (65% weighting for WM7 for one APP? that happens to screw AMD? used in webmark2001 and sysmark2001). Anyway here's my previous post at HARDOCP.
FOLLOW THIS DATA PEOPLE:http://www.anandtech.com/cpu/showdoc.html?i=1543&p=6
QUOTE: Enabling SSE gives the Athlon XP an incredible 18% increase in performance. This is a real world performance increase courtesy of SSE. [ internet content creation portion of Sysmark 2001]
Overall Sysmark from same Anandtech.com page
QUOTE:"The overall performance of the Athlon XP is improved by a little over 8% because we enabled SSE.
You can clearly see that enabling SSE paints a much brighter picture for the Athlon XP; arguably a much truer picture of its performance."
[did you guys get that much truer picture part? Note ONLY WME7(the problem app) can't see SSE unless patched, the rest of the SSE capable apps in here already id the SSE in athlonXP just fine if they are capable of SSE]
Note Intel (of course, and Intel's right hand Bapco) are ticked about this. Though Intel had no problem competely changing/recompiling code for SSE2 in SpecViewPerf.
Same Anandtech article few pages later about SpecViewPerf 6.1.2 DX-6 performance:
QUOTE:
"The Pentium 4 is much more competitive in this viewset but more importantly is the huge performance increase that is courtesy of the Palomino core enhancements. You can see that the Athlon 1.4GHz is clearly the worst performer in this benchmark, but the Palomino core enhancements give the Athlon XP 1.4GHz processor a 34% advantage over its older brother." [ While a bit (likely 2-10%, if old comments on AthlonMP in previous articles are correct) is coming from the TLB enhancements, most of the 34% is coming from SSE. Remember the comparison is same speed athlon vs athlon XP: You also see in benchmarks that don't use SSE athlon XP at same speed as Tbird 1.4 only nets a scant 2-10% or so...core enhancements other than SSE aren't worth the huge gains seen in some benchmarks ]
[ Nothing to do with SSE here, but I couldn't resist throwing in a quote from Techreport.com about "the internet is faster on P4" crap Kyle chants (webmark is awesom...HONEST...LOL) ]
QUOTE:Before the Intel fanboys start hooting, they should pause to consider Intel's own use of loopy marketing terms. NetBurst, anyone? That one is chosen to correspond with Intel's whole "makes the Internet go faster" line, which is a load of bunk. [That can be found here: http://www.tech-report.com/reviews/2001q4/athlonxp/index.x?pg=4 ]
TechReport.coms Duron 1ghz (morgan with SSE) review, check out the SpecViewPerf scores of a 1ghz Duron regularly beating a larger cached Tbird 1ghz: SSE anyone (we know this benchmark is optimized for both SSE and SSE2?
http://www.tech-report.com/reviews/2001q3/duron-1ghz/index.x?pg=7
Check out AcesHardware.com's piece about video2000 and SSE, when a MORGAN 1ghz (38.1) tramples a Tbird 1ghz (29.8) when encoding mpeg2 videos. Link: http://www.aceshardware.com/Spades/read.php?article_id=45000210
Here's the QUOTE: "PCT reports that about 15% of of the instruction mix consists of MMX instructions. However, SSE-instructions are also reported as "MMX." If you take a look at a previous MPEG2-encoding benchmark , you will notice that a Pentium III 800 performs like an Athlon 1000. This indicates that many of those MMX instructions are, in fact, SSE instructions. The Morgan on the MSI KT266 Pro performed much better than on the ABIT KT7A because of the higher memory bandwidth (DDR), but more importantly, because the DVD2000 benchmark is SSE optimized." [More from same page] "Keeping Morgan's MPEG-2 performance in mind, we can easily predict that the upcoming desktop Palomino will take the top spot in such SSE (non SSE-2) optimized applications." [ LOOK Ma- No patch required for SSE! But, I'll find some way to turn it off...I don't want all that FREE PERFORMANCE]
Photoshop at VansHardware.com (Intel Benchmarketing 101 article : bapco includes HEAVY SSE, but EXCLUDES ANY 3DNOW): Note that nothing has to be done to get SSE in photoshop either! Uh, real world performance at work AGAIN.
QUOTE:"Despite its 200 MHz handicap, the 1.2 GHz Palomino easily defeated its faster clocked older brother. Although the Palomino also supports hardware prefetching, the performance difference is almost entirely due to SSE since the SysMark Photoshop test is relatively bandwidth insensitive as can be seen in this old chipset review.
Our results demonstrate that the 1.2 GHz Palomino has nearly a 30% better score than the 1.4 GHz Thunderbird in the SysMark 2000 Photoshop test.
Last year, this SSE advantage would belong solely to Intel's Pentium III. VansHardware.com LINK:
http://www.vanshardware.com/articles/2001/august/010814_Intel_SysMark/010814_Intel_SysMark.htm"
[SSE allows 200mhz slower AthlonMP 1.2 to SMOKE a Tbird 1.4!...did I mention you don't need a patch? It works in the real world ]
[Extremetech.com AthlonXP article (Loyd Case) MpegPro2 scores cut in half by 1.53XP over 1.4 Tbird! ] Quote: "MPPRO2 has been enhanced for SSE, and it seems that the Athlon XP with its new SSE capability, takes good advantage of those optimizations"
[ Well, it couldn't be that SSE crap working in the REAL WORLD again could it? Oh, and without a patch for gods sake! NO WAY!] LINK:http://www.extremetech.com/article/0,3396,apn=11&s=1005&a=16008&app=9&ap=10,00.asp
[ Same article at extremetech.com about MusicMatch 6.1 with fullblown SSE! ]
QUOTE: "Again, the Athlon XP scores big over the Pentium 4, taking advantage of the SSE enhancements built into MusicMatch 6.1, cutting the P4's time by nearly half."
[ The 1.53XP scores well over 100% faster than the Tbird 1.4! I know, this SSE stuff doesn't work in the REAL WORLD right Kyle? Did I forget to mention it NEEDS NO PATCH? ]
Since I'm tired of this and I FULLY think my point was proven, I'll quit collecting data now. All we need to know now is why Kyle insists on attempting to DUPE his readers? I'm totally open to KYLE's RESPONSE HERE. No I don't want your Intel BS. Give me something that you can PROVE. Like SSE not working in a patched WME7 after your done benchmarking...LOL. Or a list of apps that won't work with AthlonXP's SSE. Do anything, but PROVE SOMETHING. I only see you telling people 'i stand by my bs, even if I can't prove any of it'. KYLE: "Giving people the wrong expectations of current software is simply wrong and I will stand my ground on this one." ....[I guess you better look beneath your feet KYLE. Your WALKING ON AIR pal. In order to "stand you ground" you actually have to have some GROUND to stand on don't you? ]
JOJIT: Is this enough data for you? I don't believe I have to KNOW a person [personally] to know that the data the person is putting forth is an outright LIE. Wait a minute. He had NO DATA. Also, since ID software never released the AMD K62 patch for Quake2 (AMD DID IT) why did everyone use it in benchmarks after it was released? Is it NOT enough that microsoft helped fix it? I actually believe this was a condition microsoft/amd agreed upon. Something to help SELL WinXP and AthlonXP together. AMD is marketing AthlonXP as the best thing since sliced bread with WinXP. Microsoft will make the patch incorporated into WinXP's WME (probably already did). Just a guess here.
However, NONE of my data above is GUESSING. The numbers DON'T LIE. SSE is used in the REAL WORLD on AthlonXP. As such, using the patch in WME7 is giving an ACCURATE representation of performance when using AthlonXP. I might think differently if it was some kind of BETA thing that wouldn't run on some PC's, or it could crash Win2K causing you to reformat (visions of Roxio Easy CD Platinum wiping out hard drives everywhere here..ROFL). However this is JUST an SSE switch for that which is ALREADY built into the software. Has NO changes to the code. It will crash NO PC. Harms nothing.
Any proof from KYLE would be great. I'd love to take back all I said. I really liked this site. Right now I'm rethinking this, as should all of you! I will retract all LIE words if Kyle can prove to us his RESPONSE to ENKAFAN was accurate.
ENKAFAN asked if the patch ONLY improves the benchmark. To which KYLE replied: Basically YES, at this time.
Blatant LIE. Which I proved MORE than adequately above. In order for him to be even remotely right (even just regarding Windows Media Encoder), SSE would have to die on Athlon as soon as you quit benchmarking. This is again, UNTRUE. The patch will work forever after you do it. Also it ONLY applies to the one APP [WME7] that happens to SKEW 2 benchmarks (webmark 2001, and Sysmark 2001) totally in Intel's favor. The rest of the REAL WORLD, ID's a CPU's SSE BIT, and NOT (like WME7) as the CPUID string. As they say: Proof is in the pudding. Wheres your proof KYLE?
Shibumi
PS:KYLE BANNED ME AFTER THIS BUT DIDN'T RESPOND TO THE DATA. Even said I could get sued for this...ROFL. For telling the truth? Is that a crime today? I digress.