⚙️ CORE ENGINE & OVERCLOCKING CPU: Intel Core i5-3570K (Ivy Bridge, 22nm) - Golden Sample Tier. Clock Speed: 4.40 GHz ROCK SOLID (Ratio 44x) – Zero-Throttling "Concrete Wall" Stability. Voltage (VCore): 1.250V Fixed (Manual Override Mode) – Extreme Efficiency/Stability Ratio. Motherboard: MSI Z77 MPOWER (Big Bang Series). Military Class III / 16-Phase Hybrid Digital Power Design / OC Certified. PSU: XFX Pro 550W (Seasonic OEM Core) – High-stability rail. ⚡ MEMORY & GRAPHICS RAM: 32GB (4x8GB) Patriot Viper "Black Mamba" DDR3 @ 1867 MHz (XMP Profile). GPU: NVIDIA GeForce GTX 1660 SUPER (Idle Temp: 27°C). ❄️ THERMALS & CHASSIS ("THE WIND TUNNEL") Case: Antec Twelve Hundred (V3) – Full Tower Airflow Monster. CPU Cooler: Hummer Series Tower (275mm Total Height) – 4x Direct Contact Copper Heatpipes. Chassis Fans (6-Fan Extreme Airflow): Side: ARCTIC P12 Slim PWM PST (Blasting direct air on MPOWER VRMs). Top: 1x 200mm "Big Boy" Exhaust (Negative pressure king). Storage: Samsung 840 EVO 120GB (Running @ 17°C – Airflow Verified). Ambient Delta: 10°C Ambient (Winter Benchmarking Environment). Current Thermals: 38°C - 44°C (Idle/Low Load @ 4.4 GHz Locked). 🐧 LINUX KERNEL & LOW-LATENCY OVERRIDES OS: GNU/Linux (Rolling Release, March 2026). Kernel: Liquorix 6.19.8-2 (Zen-Optimized Interactive Scheduler). Extreme Boot Flags (GRUB): intel_pstate=disable, cpufreq.off=1, processor.max_cstate=0, idle=poll. Strategy: Total bypass of frequency scaling and power-saving C-States. Zero Wake-up Latency. 🛑 LATENCY DOMINATION (vs MODERN ARCHITECTURES) Cyclictest Result: < 5ms. Worst-Case Latency. The Reality Check: The "Ice Tractor" responds faster than a Core i9-14900K (Stock) running standard power management. While modern chips are "sleeping" to save pennies on the bill, this Ivy Bridge is already through the finish line. 📊 BENCHMARK PERFORMANCE: SILVERBENCH (2422 SCORE) 💀 THE "HALL OF SHAME" (STOCK TARGETS DESTROYED): Intel Targets: Core i7-6700 (Skylake): Humiliated. Your 8 threads mean nothing against the 16-Phase MPOWER stability and 4.4 GHz raw clock. Core i5-9400F (Coffee Lake): Beaten in responsiveness. The 9400F is a software-shackled dog; the Tractor is a pure silicon beast. Core i3-10100/10105: Surpassed. Lower memory latency and superior cycle management. Intel N100 (2023): A mobile toy. Completely obliterated by 2012 engineering. AMD Targets: Ryzen 5 1600 / 2600: Ivy Bridge IPC at 4.4 GHz makes Zen/Zen+ look like a slideshow in single-threaded tasks. Ryzen 7 2700: 8 cores watching the taillights of a 2012 quad-core. Kernel latency is embarrassing on Stock Zen+ compared to this setup. AMD Athlon Gold 7220U (2023): Modern architecture, budget performance. Easy prey for the Tractor. STATUS: ROCK SOLID. "The MSI Z77 MPOWER isn't a motherboard; it's a license to kill. 2012 Architecture, 2026 Tuning. Power saving is for the weak. 1.250V of pure, undiluted performance. No mercy for modern silicon."

Edited March 30Mar 30 by TheFyxxxer

Evidence of the "Ice Tractor" performance:

• System Latency: 1.86ms (Max) on Sysbench – Optimized with Liquorix Kernel & ZEN Scheduler.

• 7-Zip Benchmark: ~24,400 MIPS (Compression) / 17,700 MIPS (Decompression).

• Clock: Solid 4.4GHz on i5-3570K / MSI Z77 MPOWER

This Ivy Bridge setup is built for low-latency stability, not just raw clock speed.

Wild rig man!!! Keep up the good work

LATENCY COMPARISON: THE "ICE TRACTOR" PROJECT

1. i5-3570K "ICE TRACTOR" (Our Optimized Build)

• Clock Speed: 4.40 GHz (Locked / Zero Throttling)

• Power Management: DISABLED (C-State 0 / Idle=Poll)

• Kernel: Liquorix 6.19 (Zen-Optimized / PREEMPT)

• Min Latency: 1.03 microseconds

• Average Latency: 1.84 microseconds

• Max Latency (Jitter): 3.09 microseconds

• Wake-up Penalty: ZERO (CPU is always hot/active)

• Result: Deterministic Real-Time Response.

2. INTEL i9-14900K (Modern Stock Architecture)

• Clock Speed: Dynamic Boost (Up to 6.0 GHz)

• Power Management: ENABLED (C-States / E-Cores active)

• Kernel: Standard OS (Windows/Linux Generic)

• Min Latency: ~2.50 microseconds

• Average Latency: ~12.00 microseconds

• Max Latency (Jitter): 80.00+ microseconds (Random Spikes)

• Wake-up Penalty: Significant (Latency caused by C-State transitions)

• Result: High peak power, but inconsistent responsiveness.

3. i5-3570K (Standard Non-Optimized Setup)

• Clock Speed: Stock (3.4 - 3.8 GHz)

• Power Management: ENABLED (Balanced Mode)

• Kernel: Generic Linux Kernel

• Min Latency: ~15.00 microseconds

• Average Latency: ~45.00 microseconds

• Max Latency (Jitter): 250.00+ microseconds

• Wake-up Penalty: High

• Result: Standard legacy performance.

TECHNICAL CONCLUSION FOR HWBOT:
Modern chips chase benchmark scores but suffer from "micro-stutter" due to aggressive power-saving and complex P/E core scheduling. The "Ice Tractor" (Ivy Bridge 2012) on the MSI Z77 MPOWER platform bypasses 14 years of efficiency bloat. By forcing the CPU into an "Always-Hot" state (1.250V fixed) and using extreme kernel flags (idle=poll), we achieved a sub-5 microsecond worst-case latency. It’s a surgical tool designed for instantaneous execution, proving that 2012 engineering, when pushed to the limit, can still dominate modern silicon in pure responsiveness

TECHNICAL CONCLUSION & PHILOSOPHY (READ THIS, HWBOT)

This build is a middle finger to modern "Aquarium PC" trends. While you kids spend hundreds of dollars on useless RGB strips and tempered glass cases to watch your CPU struggle with bloatware, we are running 2012 military-grade hardware at 1.250V fixed. This Ivy Bridge isn't a museum piece; it's a surgical tool that humiliates your modern silicon in pure responsiveness.

Let’s be honest: after Windows XP Black Modded and Windows 7 Ultimate Pro 64-bit, Microsoft has only produced piles of garbage. Modern Windows versions are nothing but RAM-eating zombies, bloated with telemetry and background services that destroy latency. Linux is the only choice for those who actually know how to control their hardware.

The "Ice Tractor" bypasses 14 years of power-efficiency bloat. We don't care about your Cinebench score or your pretty lights. We care about zero-latency execution. This setup proves that pure engineering beats marketing every single day.

Stop buying toys. Start tuning your silicon.

The "Antec Twelve Hundred" Custom Cooling:
To keep the VRMs and the liquid cooling loop from melting under the constant voltage, I’ve modded a legendary Antec 1200 case:

• Front Intake: Twin DeepCool TF140S fans (Gamer Storm series) mounted on the radiator. Their double-layer blade design provides the high static pressure needed to force air through the fins.

• Side Panel Mod: A dedicated high-pressure fan targeting the VRMs and socket area. Essential for maintaining stability at 1.250V fixed, dropping power phase temps by a significant margin.

Philosophy: Total control over airflow. Brutal cooling for a brutal overclock.

Edited March 27Mar 27 by TheFyxxxer

29 minutes ago, TheFyxxxer said:

M

Edited March 27Mar 27 by TheFyxxxer

2 minutes ago, TheFyxxxer said:

ANTEC BIG BOY 200mm MAXIMUM AIR OUTFLOW

THE LICENSE TO KILL...LOL

"The Core of the Ice Tractor: Achieving System Determinism"

"While most overclocking focuses on reaching the highest peak frequency (Max MHz) for a fleeting second, Project The Ice Tractor is an exercise in System Determinism.

What you see in my benchmarks isn't just 'speed'; it's mathematical certainty. By combining the 'Concrete Wall' physical shielding with a Liquorix Kernel and the ZEN scheduler, I have eliminated the 'noise' that plagues modern PC architectures.

How I achieve Determinism:

1. Hardware Isolation: Moving audio to an external Vention 7.1 controller isn't just about sound; it's about stripping the CPU of unpredictable IRQ interrupts. The MSI Z77 MPOWER is now a pure calculation engine, free from EMI and onboard jitter.

2. Structural Inertia: The concrete and reinforced airflow aren't for show. They provide a vibration-free environment where thermal fluctuations are non-existent, ensuring the electrical signal remains rock-solid.

3. Software Precision: My sub-5µs latency isn't a 'peak' value—it's a constant. Under Linux, the system operates with near real-time precision.

In short: I haven't just overclocked a 3570K; I have turned it into a Deterministic Fortress. It doesn't just run fast—it runs exactly the same, every single cycle, forever. This is Engineering, not just Benchmarking.

"The Ultimate Proof: Prime95 5K Stability"

"To those who doubt the 'Concrete Wall' (Muro di Cemento): this system isn't just for show. It has survived the ultimate torture: Prime95 5K FFTs at only 73°

On a 2012 Ivy Bridge platform, passing the 5K test is a declaration of war against instability. This specific load hammers the L1/L2 caches and puts the Military Class III VRMs under extreme electrical pressure.

Why it matters for my work:
I use this machine daily with my Roland DJ-505. In live audio, even a micro-fluctuation in voltage can cause a 'pop' or a latency spike. My 'Tractor' handles the 5K stress test with the same ease it handles a 4-hour DJ set.

System Hardening Summary:

• Vention 7.1 External Audio: Zero onboard IRQ noise for the CPU.

• Extreme Airflow: Zero thermal drift during the 5K torture.

• Liquorix Kernel/ZEN: Zero jitter, sub-5µs latency at all times.

I don't just 'overclock'. I build indestructible systems. If your PC can't pass a 5K test while maintaining surgical latency, you haven't built a 'Muro de Cemento! you've built a house of cards."

https://browser.geekbench.com/v6/cpu/17366487

https://browser.geekbench.com/v6/cpu/17366177

Forget brute force for a second. Look at the Sysbench Latency on the Ice Tractor. 1.84ms average latency with a 95th percentile of 1.82ms. Running on Linux Mint 22.3 with Liquorix Kernel. This is why I built the Concrete Wall: absolute thermal stability and sub-2ms response times. Can your 2026 rig match this consistency?

The silence in this thread is interesting. Perhaps it's hard to discuss actual engineering when you're used to "brute force" modern silicon?

I’ve just finalized the Sysbench CPU and Latency tests on The Ice Tractor, and the numbers should make some of you rethink your "modern" setups:

• Average Latency: 1.84ms (Constant)

• 95th Percentile: 1.82ms

• Thermal Stability: 42°C @ 4.4GHz (Rock solid on the Concrete Wall)

• OS: Linux Mint 22.3 with Liquorix Kernel (6.19-1-liquorix)

Why this matters:
While most 2026 rigs are fighting DPC latency spikes and thermal throttling on 3nm/5nm chips, this 14-year-old i5-3570K on a Z77 MPOWER is delivering surgical precision. The "Concrete Wall" isn't just a support; it's a massive thermal sink that provides a flat-line temperature curve no AIO can match.

The Challenge:
Does anyone here have a 2024-2026 system that can show a more consistent latency graph or better thermal inertia under load? My 839 Single-Core (GB6) might not be the highest, but in terms of signal purity and real-time response, this "tractor" is a beast.

Stop staring at the concrete and start looking at the logs. Who’s ready to benchmark latency against an Ivy Bridge legend?

Who let ChatGPT create a forum account ?

1 hour ago, Paul7347 said:

Who let ChatGPT create a forum account ?

chatgpt is stupid and limited... mine speaks the same but it's indescribable lol. So besides talking, can you say anything more concrete?

I can say more concrete things, you haven't said anything about tuning RAM ?

Is that not relevant despite story most of the information your CPU has to access ? You're giving up on quite the latency headroom by just running XMP there, bet it introduces some crackling in your audio

The build is old-school, I like the style it looks cool. It's an affordable base that you tuned, that's great and exactly what overclocking is at the root.

But it's got nothing on modern systems, let alone "a surgical tool that humiliates your modern silicon in pure responsiveness.". At least up until Rocket Lake, I do agree we get more and more interconnects everywhere that tend to hurt latency but that's where we're heading.

When your comparing to a 14900ks why not compare to a tuned one? Since in the data you stated above it was better at stock than a stock 3570k? Also in the comparison your using microseconds instead of milliseconds that you use elsewhere? You also don't state how your running sysbench which makes it hard to compare to.

Here's my guess at a comparison with my 7960x in my arch install with stock cpu + slight mem oc with rdimms:

❯ sysbench cpu run --threads=1 
sysbench 1.0.20 (using system LuaJIT 2.1.1720049189) 
 
Running the test with following options: 
Number of threads: 1 
Initializing random number generator from current time 
 
 
Prime numbers limit: 10000 
 
Initializing worker threads... 
 
Threads started! 
 
CPU speed: 
    events per second:  1197.31 
 
General statistics: 
    total time:                          10.0009s 
    total number of events:              11976 
 
Latency (ms): 
         min:                                    0.83 
         avg:                                    0.83 
         max:                                    2.73 
         95th percentile:                        0.84 
         sum:                                 9999.11 
 
Threads fairness: 
    events (avg/stddev):           11976.0000/0.00 
    execution time (avg/stddev):   9.9991/0.00

That seems a lot better to me than your results? Or are you running a different sysbench test? You don't state what you run so how are we supposed to compare?

I'm not sure how well sysbench captures any potential issues with audio latency. If that's all you care about then a ivy bridge i5 is going to probably be enough for your use case. Your examining a different metric that a lot of the benchmarks on hwbot, even then modern uarch has benefits in cache size per core at multiple cache levels, further improvements in limiting the downsides of speculative execution failures, better branch perdiction, larger TLB cache.... there's lots of reasons it can be better in different latency sensitive tasks then something like ivy, you just have to test it in multiple different ways.

WOW! i used sysbench cpu terminal. The result appears in microseconds because Sysbench, in the "Latency" section, measures the time taken to complete each individual event (in this case, the calculation of a series of prime numbers). The difference in the unit of measurement depends on the version of Sysbench used (its 1.0.20 vs. my older one). To run an identical and comparable test, i should rerun the command, forcing a single thread . I have an average latency of 1846.51. Converted to milliseconds, it's 1.84 ms. You're comparing an i9-7960X (16 cores) with an i5-3570K (4 cores) on different versions of Sysbench. My latency is in microseconds because I'm using version 0.4.12, yours is in milliseconds (ms) because you're using 1.0.20. Even though your single-core latency is lower (obviously, it's an i9), my 2166 total events/sec exceed your 1197 because you've locked your processor to a single thread. Run it with --threads=32 and see if your rig can handle the heat!

❯ sysbench cpu run --threads=32 
sysbench 1.0.20 (using system LuaJIT 2.1.1720049189) 
 
Running the test with following options: 
Number of threads: 32 
Initializing random number generator from current time 
 
 
Prime numbers limit: 10000 
 
Initializing worker threads... 
 
Threads started! 
 
CPU speed: 
    events per second: 29364.17 
 
General statistics: 
    total time:                          10.0009s 
    total number of events:              293711 
 
Latency (ms): 
         min:                                    0.90 
         avg:                                    1.09 
         max:                                   22.91 
         95th percentile:                        1.10 
         sum:                               319933.45 
 
Threads fairness: 
    events (avg/stddev):           9178.4688/198.03 
    execution time (avg/stddev):   9.9979/0.00 
 
❯ sysbench cpu run --threads=4 
sysbench 1.0.20 (using system LuaJIT 2.1.1720049189) 
 
Running the test with following options: 
Number of threads: 4 
Initializing random number generator from current time 
 
 
Prime numbers limit: 10000 
 
Initializing worker threads... 
 
Threads started! 
 
CPU speed: 
    events per second:  4748.25 
 
General statistics: 
    total time:                          10.0008s 
    total number of events:              47494 
 
Latency (ms): 
         min:                                    0.83 
         avg:                                    0.84 
         max:                                    2.72 
         95th percentile:                        0.90 
         sum:                                39993.28 
 
Threads fairness: 
    events (avg/stddev):           11873.5000/26.62 
    execution time (avg/stddev):   9.9983/0.00

There's some 32 thread and 4 thread runs.

The screenshot you showed above of sysbench is showing the latency unit as (ms)? Are you using two different versions?

Edited March 30Mar 30 by exaberries

Impressive power, 29k EPS is out of this world! But look at the maximum latency: 22.91 ms is a huge spike compared to the average of 1.09 ms. This means that with 32 threads, the Skylake-X is starting to suffer from jitter or heat.
My 'Ice Tractor' (i5-3570K) may do fewer calculations, but my maximum latency is much closer to the average: a sign of rock-solid stability :)

Here's the bit from your screenshot showing how the unit is milliseconds not microseconds.

8 hours ago, Paul7347 said:

The build is old-school, I like the style it looks cool. It's an affordable base that you tuned, that's great and exactly what overclocking is at the root.

But it's got nothing on modern systems, let alone "a surgical tool that humiliates your modern silicon in pure responsiveness.". At least up until Rocket Lake, I do agree we get more and more interconnects everywhere that tend to hurt latency but that's where we're heading.

4 thread

Looking at your single-thread result, you reached 1197.33 eps. My 2012 i5-3570K @ 4.4GHz is pushing 1399.29 eps. That's a 17% performance gap in my favor, with much better latency too (0.71ms vs 0.83ms).

It's impressive that you need 16 cores to stay relevant, but when it comes to raw engine power, the Concrete Wall doesn't budge. My 'old' Ivy Bridge is officially outperforming your i9 in efficiency.

You’re basically driving a huge 16-wheeler truck that gets overtaken by a tuned 2012 Tractor every time there's a real sprint. Brute force is easy, but core efficiency and optimization are what the Concrete Wall is all about!

Maybe you should try to actually overclock those 16 cores, or just admit that Ivy Bridge still has more 'soul' than Skylake-X! 😉

sorry for getting confused with micro seconds

%                                                                                                                                                                                                                                       ╭─   ~ ···········································································❯ sysbench cpu run --threads=32
sysbench 1.0.20 (using system LuaJIT 2.1.1720049189)

Running the test with following options:
Number of threads: 32
Initializing random number generator from current time


Prime numbers limit: 10000

Initializing worker threads...

Threads started!

CPU speed:
    events per second: 38969.69

General statistics:
    total time:                          10.0007s
    total number of events:              389768

Latency (ms):
         min:                                    0.74
         avg:                                    0.82
         max:                                    8.34
         95th percentile:                        0.84
         sum:                               319937.27

Threads fairness:
    events (avg/stddev):           12180.2500/35.31
    execution time (avg/stddev):   9.9980/0.00

There you go if you want a slight oc, but I'll be disabling that since I really don't need it for my daily work....

And any 6700k/7700k/9900k/10900k/11900k would easily be able to do even better....

@Paul7347 don't worry I think we can defend x299's soul from ivy

Edited March 30Mar 30 by exaberries