Gideon
Golden Member
- Nov 27, 2007
- 1,531
- 3,230
- 136
Yeah, I was just about to say: "If you think you're afraid you've got news but MLID is the only source, you have nothing to worry about"
Imho he has no real clue about the gains, I would agree ADL-S will be something big but saying like there is a 20% IPC improvement is just a guess or speculation at the moment. Initially he said 10-20% and may be he saw the Geekbench leak and thought it could be bigger than usual.
Intel is listing a new Comet Lake-S processor, Core i9 10900KS
www.guru3d.com
Intel's MIGHTY 14nm should be written in semiconductor history because of its particularity. /facepalm/
Intel is listing a new Comet Lake-S processor, Core i9 10900KS
Um yeah, a little unusual it is. That would be to release a Comet Lake-S processor so shortly before the new 11th Gen processor release. But it's listed at Intel, the SKU would be a Core i9 10900KS....
Intel's MIGHTY 14nm should be written in semiconductor history because of its particularity. /facepalm/
i want to make a joke about how Intel stretched out 14nm, but the punchline would get me banned. 
For those of you interested in IPC calculated from Anandtech reviews read on.
For the rest of you, move along, don't waste your time here.
I went back through all of the Anandtech reviews and compiled the following core-to-core comparisons. In some cases minor normalization of clocks was required, but it was minimal, the worst case was 100MHz.
Except for Skylake to Sunny Cove all comparisons were desktop-to-desktop. For Skylake to Sunny Cove I compared 1065G7 to 8550u.
The MOST objective part of this was I had in some cases where I didn't just average all of the benchmarks I had to pick some to use for the final average. In my original spreadsheet they are represented by bold font but in the copy/paste that doesn't come through. If you are interested in the original spreadsheet PM me.
I tried to average benches so they were unfairly weighted to a particular set of benchmarks.
I have the least faith in the Skylake to Sunny Cove comparison due to the variation in clocks during benchmarks due to thermals. I tried to stay with Dell laptops in hopes their thermal solutions would be comparable. When Ian tests Rocket Lake we'll know the true IPC increase from Skylake to Sunny Cove.
Haswell doesn't look as good as it should because there were actually architectural improvements in Ivy Bridge, which mitigate the Haswell improvement from Sandy Bridge. Sandy to Haswell would actually be more like 16% if Ivy was a normal die shrink without architectural changes. This seems about right as Haswell was a big update.
Everybody will have their opinion on what bench/application is most relevant. Having done this work myself, and using a verified source I trust (Anandtech) I'm going with these numbers for my future reference.
Here are my final results with the data below.
For the rest of you, move along, don't waste your time here.
I went back through all of the Anandtech reviews and compiled the following core-to-core comparisons. In some cases minor normalization of clocks was required, but it was minimal, the worst case was 100MHz.
Except for Skylake to Sunny Cove all comparisons were desktop-to-desktop. For Skylake to Sunny Cove I compared 1065G7 to 8550u.
The MOST objective part of this was I had in some cases where I didn't just average all of the benchmarks I had to pick some to use for the final average. In my original spreadsheet they are represented by bold font but in the copy/paste that doesn't come through. If you are interested in the original spreadsheet PM me.
I tried to average benches so they were unfairly weighted to a particular set of benchmarks.
I have the least faith in the Skylake to Sunny Cove comparison due to the variation in clocks during benchmarks due to thermals. I tried to stay with Dell laptops in hopes their thermal solutions would be comparable. When Ian tests Rocket Lake we'll know the true IPC increase from Skylake to Sunny Cove.
Haswell doesn't look as good as it should because there were actually architectural improvements in Ivy Bridge, which mitigate the Haswell improvement from Sandy Bridge. Sandy to Haswell would actually be more like 16% if Ivy was a normal die shrink without architectural changes. This seems about right as Haswell was a big update.
Everybody will have their opinion on what bench/application is most relevant. Having done this work myself, and using a verified source I trust (Anandtech) I'm going with these numbers for my future reference.
Here are my final results with the data below.
P4 to Conroe | 82.7% |
Conroe to Nahalem | 21.2% |
Nahalem to Sandy Bridge | 13.7% |
Sandy Bridge to Ivy Bridge | 6.8% |
Ivy Bridge to Haswell | 8.9% |
Haswell to Skylake | 9.5% |
Skylake to Sunny Cove | 21.3% |
| Clock | 2.93 | 3 | ||
| Sysmark 2004 | X6800 | Pentium D 930 | Normalized | IPC |
| Overall | 371 | 211 | 206 | 80.0% |
| Internet Content Creation | 482 | 256 | 250 | 92.8% |
| Office Productivity | 285 | 174 | 170 | 67.7% |
| 3D Content Creation | 447 | 232 | 227 | 97.3% |
| 2D Content Creation | 568 | 302 | 295 | 92.6% |
| Web Publication | 442 | 240 | 234 | 88.6% |
| Communication And Networking | 202 | 142 | 139 | 45.7% |
| Document Creation | 380 | 206 | 201 | 88.9% |
| Data Analysis | 302 | 180 | 176 | 71.8% |
| Average | 80.6% | |||
| PC WorldBench 5 | ||||
| Overall WorldBench Score | 156 | 99 | 97 | 61.3% |
| 3D Rendering | ||||
| 2dsmax7 | 4.11 | 2.13 | 2 | 97.6% |
| Cinebench 1CPU | 486 | 256 | 250 | 94.4% |
| Cinebench XCPU | 892 | 460 | 449 | 98.5% |
| Average | 96.8% | |||
| Video Encoding | ||||
| Xmpeg 5.03 with DivX 6.1 | 19.4 | 12.2 | 12 | 62.8% |
| Windows Media Encoder WMV9 | 61.6 | 32.4 | 32 | 94.7% |
| QuickTime v7.1H.264 (sec to encode) | 120 | 223.2 | 229 | 86.0% |
| Average | 81.2% | |||
| Audio Encoding | ||||
| iTunes 6 MP3 | 26 | 48 | 49 | 84.6% |
| P4 to Conroe Average | 82.7% |
| Nahalem | Conroe | ||
| SYSmark 2007 | i7-965-3.2GHz | QX9770-3.2GHz | IPC |
| SysMarks | 238 | 222 | 7.2% |
| E-Learning | 208 | 198 | 5.1% |
| Video Creation | 277 | 265 | 4.5% |
| Productivity | 234 | 224 | 4.5% |
| 3D | 239 | 207 | 15.5% |
| Average | 7.3% | ||
| 3D Rendering | |||
| POV-Ray 3.7 beta 29 | 4202 | 2641 | 59.1% |
| Cinebench R10 - 1CPU | 4475 | 3937 | 13.7% |
| Cinebench R10 - XCPU | 18810 | 14065 | 33.7% |
| 3dsmax CPU Composite | 17.6 | 13.1 | 34.4% |
| Average | 35.2% | ||
| Encoding | |||
| x264 HD pass 1 | 85.8 | 73.2 | 17.2% |
| x264 HD pass 2 | 31.6 | 21.3 | 48.4% |
| DivX (seconds) | 32.8 | 42.4 | 22.6% |
| Windows Media Encoder (sec) | 24 | 29 | 20.8% |
| iTunes MP3 (seconds) | 26.4 | 25.2 | -4.5% |
| Average | 20.9% | ||
| Conroe to Nahalem | 21.2% |
| Clock | 3.4 | 3.33 | ||
| Sysmark 2007 | 2600K | i7-975 | Normalized | IPC |
| Overall | 274 | 251 | 245 | 11.8% |
| Photoshop CS4 | ||||
| Retouch Artists Benchmark (seconds) | 11.3 | 14.4 | 14 | 27.4% |
| File Compression/Decompression | ||||
| PAR2 Multithreaded (seconds) | 17.3 | 23.7 | 23 | 37.0% |
| WinRAR 3.80 Compression (sec) | 59.6 | 74 | 72 | 24.2% |
| 7-Zip Benchmark | 19744 | 20217 | 20642 | -2.3% |
| 7-Zip Compression | 4611 | 4447 | 4540 | 3.7% |
| Average | 15.6% | |||
| 3D Rendering | ||||
| Blender 3D Character Render (sec) | 40.1 | 46.8 | 46 | 16.7% |
| POV-Ray 3.7 | 4875 | 4379 | 4471 | 11.3% |
| 3dsmax 9 | 20.1 | 17.7 | 18 | 13.6% |
| Cinebench 10 1CPU | 5991 | 4651 | 4749 | 28.8% |
| Cinebench XCPU | 22875 | 20407 | 20836 | 12.1% |
| Average | 16.5% | |||
| Video Encoding | ||||
| Xmpeg+DivX Encode (seconds) | 29.1 | 31.3 | 31 | 7.6% |
| Windows Media Encoder WMV9 (sec) | 20 | 23 | 23 | 15.0% |
| x264 HD 1st Pass | 106.4 | 91.7 | 94 | 16.0% |
| x264 HD 2nd Pass | 36.3 | 33.1 | 32 | 12.0% |
| Average | 12.6% | |||
| Visial Studio 2008 (minutes) | 18.6 | 21 | 21 | 12.9% |
| Flash video - Sorenson Squeeze (sec) | 72.7 | 90.4 | 89 | 24.3% |
| Excel Math - Monte Carlo Sim (sec) | 11.1 | 12 | 12 | 8.1% |
| Excel Math Operations (seconds) | 3.4 | 3.183 | 3 | -6.4% |
| Nahalem to Sandy Bridge | 13.7% |
| Clock | 3.9 | 3.8 | ||
| SYSmark 2012 | 3770K | 2600K | Normalized | IPC |
| SysMarks Overall | 228 | 212.0 | 217.6 | 4.8% |
| Office Productivity | 189 | 176.0 | 180.6 | 4.6% |
| Media Creation | 218 | 197.0 | 202.2 | 7.8% |
| Productivity | 235 | 221.0 | 226.8 | 3.6% |
| Data/Financial Analysis | 277 | 268.0 | 275.1 | 0.7% |
| 3D Modeling | 260 | 234.0 | 240.2 | 8.3% |
| System Management | 200 | 187.0 | 191.9 | 4.2% |
| Average | 4.9% | |||
| SYSmark 2007 | ||||
| SysMarks Overall | 303 | 274.0 | 281.2 | 7.7% |
| Productivity | 276 | 283.0 | 290.4 | -5.0% |
| E-Learning | 308 | 244.0 | 250.4 | 23.0% |
| Video Creation | 293 | 255.0 | 261.7 | 12.0% |
| 3D | 340 | 318.0 | 326.4 | 4.2% |
| Average | 8.4% | |||
| 3D Rendering | ||||
| POV-Ray 3.7 beta 29 | ||||
| Cinebench R11.5 - 1CPU | 1.66 | 1.5 | 1.6 | 6.4% |
| Cinebench R11.5 - XCPU | 7.61 | 6.9 | 7.0 | 8.1% |
| 3dsmax R9 | 21.8 | 20.1 | 20.6 | 5.7% |
| Average | 6.7% | |||
| Encoding | ||||
| x264 HD pass 1 | 104.2 | 94.9 | 97.4 | 7.0% |
| x264 HD pass 2 | 41 | 36.0 | 36.9 | 11.0% |
| Average | 9.0% | |||
| Build Chromium Product Visual Studio (min) | 17.7 | 18.6 | 18.1 | 2.4% |
| Photoshop CS4 - Retouch Artist (sec) | 10.3 | 11.3 | 11.0 | 6.9% |
| Compression and Encryption | ||||
| 7-zip - 32MB Dictionary | 22810 | 19744 | 20263.6 | 12.6% |
| AES-128 - True Crypt 7.1 | 3.7 | 3.4 | 3.5 | 6.0% |
| Average | 9.3% | |||
| Sandy Bridge to Ivy Bridge | 6.8% |
| 4770K | 3770K | IPC | |
| POV-Ray 3.7 | 1541.3 | 1363.6 | 13.0% |
| Cinebench R11.5 - 1CPU | 1.78 | 1.7 | 7.2% |
| Cinebench R11.5 - XCPU | 8.07 | 7.6 | 6.0% |
| 7-zip single thread | 4807 | 4716.0 | 1.9% |
| 7-zip multithreaded | 23101 | 22810.0 | 1.3% |
| Kraken Java Script (Chrome) ms | 1218.3 | 1313 | 7.8% |
| PCMark-7 Overall | 6747 | 6268.0 | 7.6% |
| x264 HD 1st Pass | 79.1 | 74.8 | 5.7% |
| 2x64 HD 2nd Pass | 16.5 | 14.6 | 13.0% |
| TrueCrypt AES | 4.4 | 3.7 | 18.9% |
| Visual Studio 2012 - Build Firefox (minutes) | 20.1 | 23.1 | 14.9% |
| Ivy Bridge to Haswell | 8.9% |
| Clock | 3.4 | 3.33 | ||
| 6700K | 4770K | Normalized | IPC | |
| WinRAR 5.01 Compression (sec) | 48.97 | 54.5 | 51 | 3.3% |
| 7-Zip Compression | 26370 | 24100 | 25954 | 1.6% |
| 3D Particle Movement single thread | 140.7 | 129.37 | 139 | 1.0% |
| 3D Particle Movement multithread | 803.68 | 727.64 | 784 | 2.6% |
| Cinebench 10 1CPU | 9052 | 7718 | 8312 | 8.9% |
| Cinebench 10 XCPU | 36747 | 30095 | 32410 | 13.4% |
| x264 HD 1st Pass | 133.48 | 112.43 | 121 | 10.2% |
| x264 HD 2nd Pass | 55.9 | 46.7 | 50 | 11.2% |
| Google Octane v2 | 45345 | 32193 | 34669 | 30.8% |
| WebXPRT | 2949 | 2594 | 2794 | 5.6% |
| Dolphin Emulation (minutes) | 6.47 | 7.63 | 7 | 9.5% |
| Fastone Image Viewer 4,9 (seconds) | 34 | 40 | 37 | 9.2% |
| Sunspider (ms) | 127 | 121 | 112 | -11.5% |
| Mozilla Kraken 1.1 (ms) | 735 | 1091 | 1013 | 37.8% |
| Haswell to Skylake | 9.5% |
| Clock | 3.9 | 4.2 | ||
| 1065G7 | 8650u | Normalized | IPC | |
| PCMark 10 - Essentials | 9325 | 8413 | 7812 | 19.4% |
| PCMark 10 - Productivity | 7008 | 6480 | 6017 | 16.5% |
| PCMark 10 - Digital Content Creation | 3902 | 3035 | 2818 | 38.5% |
| PCMark 10 - Overall | 4546 | 3875 | 3598 | 26.3% |
| Cinebench R15 single thread | 181.14 | 170 | 158 | 14.7% |
| Cinebench R15 multithread | 826.7 | 658.84 | 612 | 35.1% |
| x264 HD 1st Pass | 73.72 | 68.81 | 64 | 15.4% |
| x264 HD 2nd Pass | 14.37 | 13.85 | 13 | 11.7% |
| Google Octane v2 | 40002 | 35532 | 32994 | 21.2% |
| WebXPRT 3 | 223 | 208 | 193 | 15.5% |
| WebXPRT 2015 | 593 | 557 | 517 | 14.7% |
| Mozilla Kraken 1.1 (ms) | 956.9 | 1123 | 1209 | 26.4% |
| Skylake to Sunny Cove | 21.3% |
Thanks for the nice data. AFAIK Intel uses a GeoMean for their IPC summaries. GeoMean is commonly used for growth rates. Might be more comparable if you would too.
dullard
Elite Member
- May 21, 2001
- 24,330
- 2,560
- 126
It helps to describe why people use geometric means. Geometric means are almost always slightly lower than averages. But when you need to eventually multiply effects, such as multiplying increases in IPC, the geometric mean yields the correct result when averages overestimate the final result. Geometric means are based on multiplication while averages are based on addition. So anything that you want to multiply effects together does better with geometric means.
For example, suppose the initial IPC was 100. Suppose there is a 30% gain, then a 10% gain, then a 20% gain. The final IPC is 100*1.3*1.1*1.2 = 171.6.
Average gain: (1.3 + 1.1 +1.2) / 3 = 1.2. Try the IPC calculation using averages: 100 * 1.2 * 1.2 * 1.2 = 172.8. This is slightly higher than the true result of 171.6.
Geometric gain: (1.3*1.1*1.2)^(1/3) = 1.1972. This 1.1972 geometric mean gain is just slightly lower than the average gain of 1.2. Try the IPC calculation using geometric mean: 100 * 1.972 * 1.972 * 1.972 = 171.6. This exactly matches the IPC gain if you used the actual numbers.
Last edited:
Sandy Bridge to Ivy Bridge wasn't 6.8%, more like 3-4% and if you round it to 5% it's fine, actually all should be rounded because depending on the apps and games IPC gains will alyways differ from test to test.
I'm not making the numbers up. They are coming directly from the Anandtech review. The limited tests he choose resulted in that number.
I'm recalculating using geomeans now...
And yes, benchmarks can fluctuate wildly. That's why I stay with Anandtech results only, and only compare numbers contained within the same review. Things can change from review-to-review with the same processor. I have lived through all of these releases and more unfortunately
I'm saying this because I remember these results being considered quite valid and almost canonical when the reviews were published.This is a "throughput" analysis of Intel CPU architecture from P4 through Sunny Cove computed from Anandtech benchmark results.
Notes:
This is not IPC as my colleagues have pointed out but a measure of the speed a processor computes a workload as compared to another at the same frequency.
Only CPU's from within the same testing review are compared.
I tried to maintain 3 significant digits because that is is how most results were reported. That being said precision and accuracy are unknown with these benchmarks.
If Ivy Bridge results seems higher than expected, keep in mind Ivy Bridge was a "Tick+" and had as Anand wrote "a more aggressive turbo" than Sandy Bridge. This could skew results.
Here are the Ivy Bridge core changes:
- Data structures previously statically shared between threads can now be dynamically shared (e.g. DSB queue), improves single threaded performance
- FP/integer divider delivers 2x throughput compared to Sandy Bridge
- MOV instructions no longer occupy an execution port, potential for improved ILP when MOVs are present
I personally have the least faith in the Skylake to Sunny Cove calculations because due to thermal throttling clocks are most likely wildly fluctuating, meaning large performance variation among the same CPU in different laptops. I compared Sunny Cove to the fastest 8550u in each test. We will have better Skylake to Sunny Cove data when RKL is released and I will update these results then.
Notes:
This is not IPC as my colleagues have pointed out but a measure of the speed a processor computes a workload as compared to another at the same frequency.
Only CPU's from within the same testing review are compared.
I tried to maintain 3 significant digits because that is is how most results were reported. That being said precision and accuracy are unknown with these benchmarks.
If Ivy Bridge results seems higher than expected, keep in mind Ivy Bridge was a "Tick+" and had as Anand wrote "a more aggressive turbo" than Sandy Bridge. This could skew results.
Here are the Ivy Bridge core changes:
- Data structures previously statically shared between threads can now be dynamically shared (e.g. DSB queue), improves single threaded performance
- FP/integer divider delivers 2x throughput compared to Sandy Bridge
- MOV instructions no longer occupy an execution port, potential for improved ILP when MOVs are present
I personally have the least faith in the Skylake to Sunny Cove calculations because due to thermal throttling clocks are most likely wildly fluctuating, meaning large performance variation among the same CPU in different laptops. I compared Sunny Cove to the fastest 8550u in each test. We will have better Skylake to Sunny Cove data when RKL is released and I will update these results then.
Intel Generational "throughput" Comparison Results | Geomean | Average | ||
| P4 to Conroe | 82.7% | 83.4% | ||
| Conroe to Nahalem | 20.2% | 22.4% | ||
| Nahalem to Sandy Bridge | 11.8% | 12.2% | ||
| Sandy Bridge to Ivy Bridge | 6.7% | 6.9% | ||
| Ivy Bridge to Haswell | 8.7% | 8.9% | ||
| Haswell to Skylake | 8.9% | 9.5% | ||
| Skylake to Sunny Cove | 21.0% | 21.3% | ||
| P4 to Conroe | 2.93 | 3 | Clock | |
| Sysmark 2004 | X6800 | Pentium D 930 | Normalized | IPC |
| Overall | 371 | 211 | 206 | 80.0% |
| Internet Content Creation | 482 | 256 | 250 | 92.8% |
| Office Productivity | 285 | 174 | 170 | 67.7% |
| 3D Content Creation | 447 | 232 | 227 | 97.3% |
| 2D Content Creation | 568 | 302 | 295 | 92.6% |
| Web Publication | 442 | 240 | 234 | 88.6% |
| Communication And Networking | 202 | 142 | 139 | 45.7% |
| Document Creation | 380 | 206 | 201 | 88.9% |
| Data Analysis | 302 | 180 | 176 | 71.8% |
| PC WorldBench 5 | ||||
| Overall WorldBench Score | 156 | 99 | 97 | 61.3% |
| 3D Rendering | ||||
| 2dsmax7 | 4.11 | 2.13 | 2 | 97.6% |
| Cinebench 1CPU | 486 | 256 | 250 | 94.4% |
| Cinebench XCPU | 892 | 460 | 449 | 98.5% |
96.8% | ||||
| Video Encoding | ||||
| Xmpeg 5.03 with DivX 6.1 | 19.4 | 12.2 | 12 | 62.8% |
| Windows Media Encoder WMV9 | 61.6 | 32.4 | 32 | 94.7% |
| QuickTime v7.1H.264 (seconds) | 0.0083 | 223.2 | 0.0044 | 90.8% |
82.8% | ||||
| Audio Encoding | ||||
| iTunes 6 MP3 (seconds) | 0.0385 | 48 | 0.0204 | 88.5% |
61.13 | 33.47 | 83.4% | ||
| P4 to Conroe (Geomean) | 82.7% | |||
| Conroe to Nahalem | Nahalem | Conroe | ||
| SYSmark 2007 | i7-965-3.2GHz | QX9770-3.2GHz | IPC | |
| SysMarks | 238 | 222 | 7.2% | |
| E-Learning | 208 | 198 | 5.1% | |
| Video Creation | 277 | 265 | 4.5% | |
| Productivity | 234 | 224 | 4.5% | |
| 3D | 239 | 207 | 15.5% | |
7.3% | ||||
| 3D Rendering | ||||
| POV-Ray 3.7 beta 29 | 4202 | 2641 | 59.1% | |
| Cinebench R10 - 1CPU | 4475 | 3937 | 13.7% | |
| Cinebench R10 - XCPU | 18810 | 14065 | 33.7% | |
| 3dsmax CPU Composite | 17.6 | 13.1 | 34.4% | |
35.2% | ||||
| Encoding | ||||
| x264 HD pass 1 | 85.8 | 73.2 | 17.2% | |
| x264 HD pass 2 | 31.6 | 21.3 | 48.4% | |
| DivX (seconds) | 0.0305 | 0.0236 | 29.3% | |
| Windows Media Encoder (seconds) | 0.0417 | 0.0345 | 20.8% | |
| iTunes MP3 (seconds) | 0.0397 | 0.0379 | 4.8% | |
50.24 | 41.80 | 24.1% | ||
22.4% | ||||
| Conroe to Nahalem (Geomean) | 20.2% |
| Nahalem to Sandy Bridge | 3.4 | 3.33 | Clock | |
| Sysmark 2007 | 2600K | i7-975 | Normalized | IPC |
| Overall | 274 | 251 | 245 | 11.8% |
| Photoshop CS4 | ||||
| Retouch Artists Benchmark (seconds) | 0.0885 | 14.4 | 0.0714 | 23.9% |
| File Compression/Decompression | ||||
| PAR2 Multithreaded (seconds) | 0.0578 | 23.7 | 0.0431 | 34.2% |
| WinRAR 3.80 Compression (seconds) | 0.0168 | 74 | 0.0138 | 21.6% |
| 7-Zip Benchmark | 19744 | 20217 | 20642 | -4.4% |
| 7-Zip Compression | 4611 | 4447 | 4540 | 1.6% |
13.2% | ||||
| 3D Rendering | ||||
| Blender 3D Character Render (seconds) | 0.0249 | 46.8 | 0.0218 | 14.3% |
| POV-Ray 3.7 | 4875 | 4379 | 4471 | 9.0% |
| 3dsmax 9 | 20.1 | 17.7 | 18 | 11.2% |
| Cinebench 10 1CPU | 5991 | 4651 | 4749 | 26.2% |
| Cinebench XCPU | 22875 | 20407 | 20836 | 9.8% |
14.1% | ||||
| Video Encoding | ||||
| Xmpeg+DivX Encode (seconds) | 0.0344 | 31.3 | 0.0326 | 5.3% |
| Windows Media Encoder WMV9 (seconds) | 0.0500 | 23 | 0.0444 | 12.6% |
| x264 HD 1st Pass | 106.4 | 91.7 | 94 | 13.6% |
| x264 HD 2nd Pass | 36.3 | 33.1 | 32 | 12.0% |
10.9% | ||||
| Flash video/Excel | ||||
| Visial Studio 2008 (minutes) | 0.0538 | 21 | 0.0486 | 10.6% |
| Flash video - Sorenson Squeeze (seconds) | 0.0138 | 90.4 | 0.0113 | 21.8% |
| Excel Math - Monte Carlo Sim (seconds) | 0.0901 | 12 | 0.0851 | 5.9% |
| Excel Math Operations (seconds) | 0.2941 | 3.183 | 0.3208 | -8.3% |
274.00 | 245.14 | 12.2% | ||
| Nahalem to Sandy Bridge | 11.8% | |||
| Sandy Bridge to Ivy Bridge | 3.9 | 3.8 | Clock | |
| SYSmark 2012 | 3770K | 2600K | Normalized | IPC |
| SysMarks Overall | 228 | 212.0 | 217.6 | 4.8% |
| Office Productivity | 189 | 176.0 | 180.6 | 4.6% |
| Media Creation | 218 | 197.0 | 202.2 | 7.8% |
| Productivity | 235 | 221.0 | 226.8 | 3.6% |
| Data/Financial Analysis | 277 | 268.0 | 275.1 | 0.7% |
| 3D Modeling | 260 | 234.0 | 240.2 | 8.3% |
| System Management | 200 | 187.0 | 191.9 | 4.2% |
4.9% | ||||
| SYSmark 2007 | ||||
| SysMarks Overall | 303 | 274.0 | 281.2 | 7.7% |
| Productivity | 276 | 283.0 | 290.4 | -5.0% |
| E-Learning | 308 | 244.0 | 250.4 | 23.0% |
| Video Creation | 293 | 255.0 | 261.7 | 12.0% |
| 3D | 340 | 318.0 | 326.4 | 4.2% |
8.4% | ||||
| 3D Rendering | ||||
| POV-Ray 3.7 beta 29 | ||||
| Cinebench R11.5 - 1CPU | 1.66 | 1.5 | 1.6 | 6.4% |
| Cinebench R11.5 - XCPU | 7.61 | 6.9 | 7.0 | 8.1% |
| 3dsmax R9 | 21.8 | 20.1 | 20.6 | 5.7% |
6.7% | ||||
| Encoding | ||||
| x264 HD pass 1 | 104.2 | 94.9 | 97.4 | 7.0% |
| x264 HD pass 2 | 41 | 36.0 | 36.9 | 11.0% |
9.0% | ||||
| Miscellaneous | ||||
| Build Chromium Product Visual Studio (minutes) | 0.0565 | 18.6 | 0.0552 | 2.4% |
| Photoshop CS4 - Retouch Artist (seconds) | 0.0971 | 11.3 | 0.0908 | 6.9% |
4.6% | ||||
| Compression and Encryption | ||||
| 7-zip - 32MB Dictionary | 22810 | 19744 | 20263.6 | 12.6% |
| AES-128 - True Crypt 7.1 | 3.7 | 3.4 | 3.5 | 6.0% |
58.14 | 54.48 | 6.9% | ||
| Sandy Bridge to Ivy Bridge (Geomean) | 6.7% |
| Ivy Bridge to Haswell | 4770K | 3770K | IPC | |
| POV-Ray 3.7 | 1541.3 | 1363.6 | 13.0% | |
| Cinebench R11.5 - 1CPU | 1.78 | 1.7 | 7.2% | |
| Cinebench R11.5 - XCPU | 8.07 | 7.6 | 6.0% | |
| 7-zip single thread | 4807 | 4716.0 | 1.9% | |
| 7-zip multithreaded | 23101 | 22810.0 | 1.3% | |
| Kraken Java Script - Chrome (ms) | 0.0008 | 0.0008 | 7.8% | |
| PCMark-7 Overall | 6747 | 6268.0 | 7.6% | |
| x264 HD 1st Pass | 79.1 | 74.8 | 5.7% | |
| 2x64 HD 2nd Pass | 16.5 | 14.6 | 13.0% | |
| TrueCrypt AES | 4.4 | 3.7 | 18.9% | |
| Visual Studio 2012 - Build Firefox (minutes) | 0.0498 | 0.0433 | 14.9% | |
26.14 | 24.04 | 8.9% | ||
| Ivy Bridge to Haswell | 8.7% | |||
| Haswell to Skylake | 3.4 | 3.33 | Clock | |
| 6700K | 4770K | Normalized | IPC | |
| WinRAR 5.01 Compression (sec) | 0.0204 | 54.5 | 0.0198 | 3.3% |
| 7-Zip Compression | 26370 | 24100 | 25954 | 1.6% |
| 3D Particle Movement single thread | 140.7 | 129.37 | 139 | 1.0% |
| 3D Particle Movement multithread | 803.68 | 727.64 | 784 | 2.6% |
| Cinebench 10 1CPU | 9052 | 7718 | 8312 | 8.9% |
| Cinebench 10 XCPU | 36747 | 30095 | 32410 | 13.4% |
| x264 HD 1st Pass | 133.48 | 112.43 | 121 | 10.2% |
| x264 HD 2nd Pass | 55.9 | 46.7 | 50 | 11.2% |
| Google Octane v2 | 45345 | 32193 | 34669 | 30.8% |
| WebXPRT | 2949 | 2594 | 2794 | 5.6% |
| Dolphin Emulation (minutes) | 0.1546 | 7.63 | 0.1411 | 9.5% |
| Fastone Image Viewer 4,9 (seconds) | 0.0294 | 40 | 0.0269 | 9.2% |
| Sunspider (ms) | 0.0079 | 121 | 0.0089 | -11.5% |
| Mozilla Kraken 1.1 (ms) | 0.0014 | 1091 | 0.0010 | 37.8% |
31.59 | 29.00 | 9.5% | ||
| Haswell to Skylake | 8.9% | |||
| Skylake to Sunny Cove | 3.9 | 4.2 | Clock | |
| 1065G7 | 8650u | Normalized | IPC | |
| PCMark 10 - Essentials | 9325 | 8413 | 7812 | 19.4% |
| PCMark 10 - Productivity | 7008 | 6480 | 6017 | 16.5% |
| PCMark 10 - Digital Content Creation | 3902 | 3035 | 2818 | 38.5% |
| PCMark 10 - Overall | 4546 | 3875 | 3598 | 26.3% |
| Cinebench R15 single thread | 181.14 | 170 | 158 | 14.7% |
| Cinebench R15 multithread | 826.7 | 658.84 | 612 | 35.1% |
| x264 HD 1st Pass | 73.72 | 68.81 | 64 | 15.4% |
| x264 HD 2nd Pass | 14.37 | 13.85 | 13 | 11.7% |
| Google Octane v2 | 40002 | 35532 | 32994 | 21.2% |
| WebXPRT 3 | 223 | 208 | 193 | 15.5% |
| WebXPRT 2015 | 593 | 557 | 517 | 14.7% |
| Mozilla Kraken 1.1 (ms) | 0.0010 | 1123 | 0.0008 | 26.4% |
316.66 | 261.68 | 21.3% | ||
| Skylake to Sunny Cove | 21.0% |
Last edited:
Solid. Only thing id mention is you're not really comparing IPC with Penryn to Nehalem because there's so many MT tests in there which are taking advantage of HT , making them more of a throughput comparison.
Technically this is a problem with comparing the others too, but they're all running a similar implementation of HT at least.
Technically this is a problem with comparing the others too, but they're all running a similar implementation of HT at least.
actually, you're not comparing IPC at all. its semantics, but you're measuring performance uplift, unless you can correlate #CPU instructions to 1 point in test X, which you obviously cannot.
furthurmore, results are not normalized for "X pts per Mhz" because not all workloads scale with frequency and all CPUs run different frequencies and different cache structures. we all know what you mean, but call it "speed up" or "improvement", just not IPC.
Captain obvious - out.
furthurmore, results are not normalized for "X pts per Mhz" because not all workloads scale with frequency and all CPUs run different frequencies and different cache structures. we all know what you mean, but call it "speed up" or "improvement", just not IPC.
Captain obvious - out.
Last edited:
I'm not making the numbers up. They are coming directly from the Anandtech review. The limited tests he choose resulted in that number.
I'm recalculating using geomeans now...
And yes, benchmarks can fluctuate wildly. That's why I stay with Anandtech results only, and only compare numbers contained within the same review. Things can change from review-to-review with the same processor. I have lived through all of these releases and more unfortunately
All other reviews have shown Ivy Bridge with IPC gains in a range of 3-4%, for example here: https://www.hardware.fr/articles/863-12/performances-frequence-egale-ddr3-2133-pci-express-3-0.html
This is the problem, never use such exact digits, they are pointless. You have to round it. For Ivy Bridge it makes sense to round it up to 5%.
Ian is betting Q5 for ADL on desktop, will give rocketlake some time......
DrMrLordX
Lifer
- Apr 27, 2000
- 20,497
- 9,575
- 136
uzzi38
Platinum Member
- Oct 16, 2019
- 2,391
- 5,025
- 116
DrMrLordX
Lifer
- Apr 27, 2000
- 20,497
- 9,575
- 136
Yeah now that I think about it. It would be funnier if BK were still there though.
Yes you are right. It is more like "throughput." Strictly speaking it's not IPC. Maybe "efficiency" as defined by compute/cycle or something like that.
Most of the tests were reported with 3 significant digits so that is generally what I carried through. I wasn't rigorous though because as you correctly state there are many problems scientifically with the precision and accuracy of benchmark results. I leave it to the reader to interpret the correct rounded result.
As for those French testing results, as I stated, I'm using Anandtech results. I checked the Ivy Bridge numbers a few times because I was surprised as well.
As reported by Anand, Ivy Bridge was designed by Intel as a "Tick+" with 5 to 15% performance gain over Sandy Bridge. Also Anand noted "more aggressive turbo" with Ivy Bridge, which could be artificially inflating the throughput numbers a bit. Here are the Ivy Bridge core changes:
- Data structures previously statically shared between threads can now be dynamically shared (e.g. DSB queue), improves single threaded performance
- FP/integer divider delivers 2x throughput compared to Sandy Bridge
- MOV instructions no longer occupy an execution port, potential for improved ILP when MOVs are present
I have added a disclaimer of sorts to the results post. Please keep in mind it's only an analysis of the Anandtech results.
Last edited:
I posted this information in the "Ryzen availability thread" but I also think it is relevant to Intel pricing so I'm posting it again here.
So this is interesting.
I noticed my Microcenter (St. Davids, PA) has 25+ 5800X's listed last night at $500.
I checked today and they are listed at $650 each.
I guess that's one way to stop the ebay reseller gouge. I guess I'd rather be gouged by Microcenter.
www.microcenter.com
Also the "out of stock" 5950X and 5900X are listed at $1000 and $800, respectively.
Ain't it funny how the market works? You can set a price but ultimately supply and demand determine pricing.
I'm thinking at these prices Microcenter will be able to keep up with demand, once the higher end parts actually get in stock.
First, are vendors allowed to price products higher than MSRP?
If other vendors start to follow suit then there could be quite a market opening for RKL at the $650 and lower price point, right?
I just noticed that the Comet Lake parts also returned to their pre-sales prices.
So this is interesting.
I noticed my Microcenter (St. Davids, PA) has 25+ 5800X's listed last night at $500.
I checked today and they are listed at $650 each.
I guess that's one way to stop the ebay reseller gouge. I guess I'd rather be gouged by Microcenter.
Eight-Core : Processors/CPUs : Micro Center
Microcenter.com has the best deals on CPUs, both Intel & AMD Processors, to start your PC build off right! Stop by and shop for one at you local Micro Center Computer Store!
Ain't it funny how the market works? You can set a price but ultimately supply and demand determine pricing.
I'm thinking at these prices Microcenter will be able to keep up with demand, once the higher end parts actually get in stock.
First, are vendors allowed to price products higher than MSRP?
If other vendors start to follow suit then there could be quite a market opening for RKL at the $650 and lower price point, right?
I just noticed that the Comet Lake parts also returned to their pre-sales prices.
Asterox
Senior member
- May 15, 2012
- 974
- 1,665
- 136
Microcenter scalpers, yes it is legit.I posted this information in the "Ryzen availability thread" but I also think it is relevant to Intel pricing so I'm posting it again here.
So this is interesting.
I noticed my Microcenter (St. Davids, PA) has 25+ 5800X's listed last night at $500.
I checked today and they are listed at $650 each.
I guess that's one way to stop the ebay reseller gouge. I guess I'd rather be gouged by Microcenter.
Also the "out of stock" 5950X and 5900X are listed at $1000 and $800, respectively.Eight-Core : Processors/CPUs : Micro Center
Microcenter.com has the best deals on CPUs, both Intel & AMD Processors, to start your PC build off right! Stop by and shop for one at you local Micro Center Computer Store!
Ain't it funny how the market works? You can set a price but ultimately supply and demand determine pricing.
I'm thinking at these prices Microcenter will be able to keep up with demand, once the higher end parts actually get in stock.
First, are vendors allowed to price products higher than MSRP?
If other vendors start to follow suit then there could be quite a market opening for RKL at the $650 and lower price point, right?
I just noticed that the Comet Lake parts also returned to their pre-sales prices.
For example Mindfactory prices in euro.
Sockel AM4
AM4 CPU | Hier sparen :: Bis 8 Kerne :: Bis 4.30GHz Turbotakt :: Für DDR4-Speicher geeignet :: AMD ++ Einfache Bestellung und schnelle Lieferung
www.mindfactory.de
Okay strange. All prices at my MC back to normal.
www.microcenter.com
Eight-Core : Processors/CPUs : Micro Center
Microcenter.com has the best deals on CPUs, both Intel & AMD Processors, to start your PC build off right! Stop by and shop for one at you local Micro Center Computer Store!
Does anyone have an idea about what percentage of multicore workloads will scale properly with assymetrical cores? I know some programs divide a workload evenly between cores but need to wait for all threads to finish before moving forward. Might this be a common problem for Alderlake?
SAAA
Senior member
- May 14, 2014
- 541
- 126
- 116
I think that should be solved on the OS front, just like multicore CPUs, SMT or even Bulldozer approach with modules (that were loaded in a poor order till they threated them as 1-2-3-4 cores + SMT).
A purely parallel load? Well, the small cores won't be that much slower in the first place, at worst I believe half the speed:
Goldmont 1.5 IPC x 5GHz = 7.5 units of work
Gracemont 1.0 IPC x 3.5 to 4 GHz = 3.5 to 4 units of work
Making asymmetrical loads scale well might be hard but if every smartphone can do this (even on 3 different clusters) I don't see Microsoft and Intel not pulling it off.
A good idea would be to run something like OS and other apps on the small cores constantly (think anti-virus, mail or background apps), then every other program will see only big cores with SMT threads.
Ideally when the load is light enough the big cores could be turned off and the workload go into the small cores clusters only, just as today cores throttle to idle clock speeds when they are not doing much.
TRENDING THREADS
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 19K
-
Solved! Speculation: Zen 4 (EPYC 4 "Genoa", Ryzen 7000, etc.)- Started by Vattila
- Replies: 12K
-
-
-
Facebook
Twitter