Very poor performance with Intel core i5-11600KF with stock 4.9 GHz

[bug]

Ok I'm sorry, but I got angry because @Assimilator reacted with "Haha" to a couple of my posts in other threads. And in this thread I just wanted to know if this kind of performance is expected from my system, and that's solved. Also I thought you can see where any forum member lives by looking at their profile (from their IP), even if they didn't set it. Like when I tried to create my account here, it automatically detected my country and said I was not allowed to register due to too much spam from my country (and I had to contact TechPowerUp and then I was able to register). But if you don't actually know where I live, I would rather not like to disclose it anyway.


Uh, but the video is still using VP9 codec, as I can see in the "stats for nerds":
View attachment 355338

Yes, but the browser decided "no more hw video decode for you". Enjoy.

 

[wNotyarD]

Hm, I forgot if anyone has suggested this already, but a last-ditch attempt mught be trying Firefox. I am fairly certain that the x264 extension does exist there too and, since it’s the only browser remaining that doesn’t run Chromium, it might have different rendering settings in its engine, including not having whatever is blocking x264 acceleration on Fermi. Worth a shot.

If this also fails, well… the only option is to upgrade. And I am sorry, but having D-Sub as a REQUIREMENT is going to fuck you over anyway - no modern GPU has it or will have it going foreward. Hell, DVI too is dead.

Although Firefox with h264ify may offload the decoding to the Fermi and alleviate his i5 usage, it wouldn't give him anything above 1080p60 on YT (which I guess wouldn't chop by using software decoding for starters).

 

[Onasi]

@wNotyarD
Probably, was just throwing it out there. Haven’t had to mess around with extensions and YT encoders for a while, so I am a bit out of date on what allows what. But yeah, I wouldn’t be surprised if YT has 4K60 streams only via VP9 and AV1.

 

[dragosxlk]

Ok I'm sorry, but I got angry because @Assimilator reacted with "Haha" to a couple of my posts in other threads. And in this thread I just wanted to know if this kind of performance is expected from my system, and that's solved. Also I thought you can see where any forum member lives by looking at their profile (from their IP), even if they didn't set it. Like when I tried to create my account here, it automatically detected my country and said I was not allowed to register due to too much spam from my country (and I had to contact TechPowerUp and then I was able to register). But if you don't actually know where I live, I would rather not like to disclose it anyway.


Uh, but the video is still using VP9 codec, as I can see in the "stats for nerds":
View attachment 355338

The video in question is still vp9 cause it is a old video and that is how was uploaded and saved , newly uploaded videos / music videos are converted into av1 by yt upload engine

 

[agent_x007]

Is money a joke to you? I wouldn't even have posted this if this was the case. I'm NOT buying anything until 2027. The stuff in my country is 3x more expensive than what you said. Also I too have told multiple times that I NEED a VGA Port. There are adapters, but I'm NOT interested in them.

You are not interested even in passive DVI-I adapter to VGA ?
Just FYI : Maxwell 2.0 is last card generation with DVI-I support (DVI-I = DVI-D + VGA), so in future just search for GTX 950s or 750s and get a passive adapter from DVI.
If GPUs are too expensive for you, sell current KF CPU and get non-F CPU (locked/unlocked, not sure if you need OC support).
I assume your MB has VGA port on the back.

 

[bug]

You are not interested even in passive DVI-I adapter to VGA ?
Just FYI : Maxwell 2.0 is last card generation with DVI-I support (DVI-I = DVI-D + VGA), so in future just search for GTX 950s or 750s and get a passive adapter from DVI.
If GPUs are too expensive for you, sell current KF CPU and get non-F CPU (locked/unlocked, not sure if you need OC support).
I assume your MB has VGA port on the back.

He's basically asking for the impossible. No card with analog outputs has the decoders for today's YouTube. It is that simple. He's adamant he needs all those features and cannot change hardware, but one day that will change. Because I'm damn sure the world won't bend over backwards to build the video card he wants. And then give it to him for free.
I think we've all been there, reading about hardware we could not afford at the time. And we all know how that turned out.

 

[Marcus L]

He's basically asking for the impossible. No card with analog outputs has the decoders for today's YouTube. It is that simple. He's adamant he needs all those features and cannot change hardware, but one day that will change. Because I'm damn sure the world won't bend over backwards to build the video card he wants. And then give it to him for free.
I think we've all been there, reading about hardware we could not afford at the time. And we all know how that turned out.

Correct, OP has stated he CAN'T afford another GPU for years neither used apparently so, also REFUSES to use an DVI-VGA adaptor even if he could get a newer GPU, so it leaves him 1 choice (unkess he can sell his i5 and swap it out for one with IGP, though am guessing this is off the cards as he can't afford a newer used GPU for $30-40) he will just have to suck up watching 1080p/60 on his 4k TV

I have a 4kTV and the only 4K content I watch on it is the odd Prime video title that has 4k support, I have basic NF which I believe is 1080p, not sure about Disney, and the kids mainly watch YT on it, I use my phone or computer for YT, tbh unless you are sitting in front of it like a monitor you will be hard pressed to notice and even so..... 1st world problems

 

[eidairaman1]

Zeb GT 730 4 GD3

Zebronics ZEB-GT730 4GD3 Graphics Card with 4GB GDDR3 features a silent fan, multiple outputs, and supports up to 4096x2160 resolution at 24Hz.

zebronics.com

You will need to remove the heatsink to see what ram is on that card and get a picture of it because our bios database does not have a GT730 with 4GB of GDDR3 like the ZEB website claims the card to have.

 

[IsHacker]

I was forced to overclock this cpu to 5.6 GHz because at stock it wasn't even able to play 4K 60FPS Youtube videos? And CPU usage was over 70% in games like Minecraft (about to bottleneck!) even with a GT 610!
Is this kind of performance expected from this CPU? Also the Cpu uses around 75W at 3% usage when idling (at stock 4.9 GHz, absolutely NO overclocking), it's really power hungry. Is that normal too?

I think I just found out why this deterioration in performance happened. It was a microcode update! Recently there was yet ANOTHER microcode update. In Linux Mint (I don't have windows anymore), I got an update for a package called "intel-microcode". This microcode update changed my UEFI resolution to 1024x768 from 1366x768. In the same way, I think a microcode update got installed through windows update back then, which reduced the performance.

 

[_roman_]

I had problems with microcode loading wth my previous intel laptop.

I decided to permanently disable any cpu microcode loading for the processor. A few years ago it was called late microcode loading in the kernel.

Usually you get the new microcode anyway with a new UEFI, or lets call it bios, mainboard update. I do not see much use to have that feature enabled.

It seems you are a beginner.

The gentoo sources, kernel.org + gentoo patchsets, have a full sections only for AMD security bugs. There are also options for intel cpus last time i was making changes in that section.

Have you ever paid attention to this, part of the output. You may pay attention to the section called bug.

bugs : sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass srso

Sienna_Cichlid /home/roman # cat /proc/cpuinfo
processor    : 0
vendor_id    : AuthenticAMD
cpu family    : 25
model        : 97
model name    : AMD Ryzen 5 7600X 6-Core Processor
stepping    : 2
microcode    : 0xa60120c
cpu MHz        : 2991.000
cache size    : 1024 KB
physical id    : 0
siblings    : 12
core id        : 0
cpu cores    : 6
apicid        : 0
initial apicid    : 0
fpu        : yes
fpu_exception    : yes
cpuid level    : 16
wp        : yes
flags        : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good amd_lbr_v2 nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cat_l3 cdp_l3 hw_pstate ssbd mba perfmon_v2 ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local avx512_bf16 clzero irperf xsaveerptr rdpru wbnoinvd cppc arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic vgif x2avic v_spec_ctrl vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid overflow_recov succor smca fsrm flush_l1d amd_lbr_pmc_freeze
bugs        : sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass srso

Those cpu bugs will increase and all will cost performance. People do not understand my motivation. I sold on purpose the ryzen 5800x. Even the second hand ryzen 3 3100 was fine for desktop use which i had for the few months transistion period.

Lessons learnt this time: It seems there is a new line with microcode in the cpuinfo output.

I do not recommend it. If you want your performance back you have to build your own kernel and set all those options yourself. I highly doubt a beginner distro based on debian or based on ubuntu is the right choice for performance.

 

[bug]

I think I just found out why this deterioration in performance happened. It was a microcode update! Recently there was yet ANOTHER microcode update. In Linux Mint (I don't have windows anymore), I got an update for a package called "intel-microcode". This microcode update changed my UEFI resolution to 1024x768 from 1366x768. In the same way, I think a microcode update got installed through windows update back then, which reduced the performance.

Unless you do before/after testing and make sure everything else is the same, you have no proof it was the microcode update.
Fwiw, I get microcode updates all the time, I've never felt any difference in performance. And I say "felt" because I bench very, very rarely.

 

[IsHacker]

Unless you do before/after testing and make sure everything else is the same, you have no proof it was the microcode update.
Fwiw, I get microcode updates all the time, I've never felt any difference in performance. And I say "felt" because I bench very, very rarely.

It couldn't be anything else. At least I don't think so.

I had problems with microcode loading wth my previous intel laptop.

I decided to permanently disable any cpu microcode loading for the processor. A few years ago it was called late microcode loading in the kernel.

Usually you get the new microcode anyway with a new UEFI, or lets call it bios, mainboard update. I do not see much use to have that feature enabled.

It seems you are a beginner.

The gentoo sources, kernel.org + gentoo patchsets, have a full sections only for AMD security bugs. There are also options for intel cpus last time i was making changes in that section.

Have you ever paid attention to this, part of the output. You may pay attention to the section called bug.

bugs : sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass srso

Sienna_Cichlid /home/roman # cat /proc/cpuinfo
processor    : 0
vendor_id    : AuthenticAMD
cpu family    : 25
model        : 97
model name    : AMD Ryzen 5 7600X 6-Core Processor
stepping    : 2
microcode    : 0xa60120c
cpu MHz        : 2991.000
cache size    : 1024 KB
physical id    : 0
siblings    : 12
core id        : 0
cpu cores    : 6
apicid        : 0
initial apicid    : 0
fpu        : yes
fpu_exception    : yes
cpuid level    : 16
wp        : yes
flags        : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good amd_lbr_v2 nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cat_l3 cdp_l3 hw_pstate ssbd mba perfmon_v2 ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local avx512_bf16 clzero irperf xsaveerptr rdpru wbnoinvd cppc arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic vgif x2avic v_spec_ctrl vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid overflow_recov succor smca fsrm flush_l1d amd_lbr_pmc_freeze
bugs        : sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass srso

Those cpu bugs will increase and all will cost performance. People do not understand my motivation. I sold on purpose the ryzen 5800x. Even the second hand ryzen 3 3100 was fine for desktop use which i had for the few months transistion period.

Lessons learnt this time: It seems there is a new line with microcode in the cpuinfo output.

I do not recommend it. If you want your performance back you have to build your own kernel and set all those options yourself. I highly doubt a beginner distro based on debian or based on ubuntu is the right choice for performance.

I am not a beginner, I've been using linux distros for years. I'm currently running linux-tkg kernel which I compiled myself. Can you elaborate more on how to set those options? Is there any difference for intel and AMD? My "bugs" section says:

bugs        : spectre_v1 spectre_v2 spec_store_bypass swapgs mmio_stale_data retbleed eibrs_pbrsb gds bhi

Thanks.

 

[_roman_]

I will emerge gentoo sources 6.15.0. I will provide you with a screenshot from the section.

Did you check the marked section?

I recommend checking every section. Like I did in 2006.

ofc you can check your current kernel, assuming you do the same bad thing to have the config.gz readable enabled.

Sienna_Cichlid /home/roman # zgrep MITIGATION* /proc/config.gz
CONFIG_CPU_MITIGATIONS=y
CONFIG_MITIGATION_PAGE_TABLE_ISOLATION=y
CONFIG_MITIGATION_RETPOLINE=y
CONFIG_MITIGATION_RETHUNK=y
CONFIG_MITIGATION_UNRET_ENTRY=y
CONFIG_MITIGATION_IBPB_ENTRY=y
CONFIG_MITIGATION_SRSO=y
CONFIG_MITIGATION_SLS=y
CONFIG_MITIGATION_RETBLEED=y
CONFIG_MITIGATION_SPECTRE_V1=y
CONFIG_MITIGATION_SPECTRE_V2=y
CONFIG_MITIGATION_SSB=y
CONFIG_ARCH_CONFIGURES_CPU_MITIGATIONS=y
# CONFIG_RESET_ATTACK_MITIGATION is not set

Thanks for your question - So sad it's somewhere else hidden. Yes maybe not the right place but i prefer all in one spot.

CONFIG_RESET_ATTACK_MITIGATION: │
│ │
│ Request that the firmware clear the contents of RAM after a reboot │
│ using the TCG Platform Reset Attack Mitigation specification. This │
│ protects against an attacker forcibly rebooting the system while it │
│ still contains secrets in RAM, booting another OS and extracting the │
│ secrets. This should only be enabled when userland is configured to │
│ clear the MemoryOverwriteRequest flag on clean shutdown after secrets │
│ have been evicted, since otherwise it will trigger even on clean │
│ reboots. │
│ │
│ Symbol: RESET_ATTACK_MITIGATION [=n] │
│ Type : bool │
│ Defined at drivers/firmware/efi/Kconfig:170 │
│ Prompt: Reset memory attack mitigation │
│ Depends on: EFI [=y] && EFI_STUB [=y] │
│ Location: │
│ -> Device Drivers │
│ -> Firmware Drivers │
│ -> EFI (Extensible Firmware Interface) Support │
│ -> Reset memory attack mitigation (RESET_ATTACK_MITIGATION [=n])

the other thing is how you compile your packages. or your distro does. Except a few instructions like AMD XOP, usually my hardware is down compatible. So i optimise for the current hardware i own. the topic about how to compile and waht to compile is out of the scope. This also has heavy impact on performance. Also how you setup your machine, filesystem and dram. this is out of scope.

compiling packages was a great teacher since 2006. A good metric on how to find performance. It's out of scope. Arch linux and maybe gentoo linux wiki may be a starting point, when you know what to search for. Or even redhat docs. Anyway out of scope of your quesiton.

-- short: build your kernel with a proper toolchain with proper compiler settings. find something to really compare
e.g. compile libreoffice from source in offline mode and compare how many seconds - 3 runs took. or similar benchmarks for you.
you should be aware what you do anyway - out of scope to go into depth.