Vnish Firmware in Production

What Our Fleet Testing Revealed on S19j Pro & S21 Pro 

At Frontier Mining, we spend a lot of time thinking about the full lifecycle of mining hardware.

Hashrate matters. Efficiency matters. But what ultimately matters most is how long a fleet remains economically competitive as network difficulty rises and power markets fluctuate.

That means our role as an operator goes well beyond simply keeping machines online. It involves continuously evaluating firmware, power profiles, and operational strategies that allow clients to extract maximum value from their hardware over time.

Today, Frontier operates more than 500 MW of active mining infrastructure across multiple sites and hardware generations. Within that footprint, over 200,000 ASIC miners have run Vnish firmware in production environments, with our engineering teams leading the fine-tuning and performance evaluation process while our 24/7 Network Operations Center (NOC) continuously monitors efficiency, stability, and fleet performance.

Over the past several years, our operations and engineering teams have conducted extensive production testing on Vnish firmware across two major hardware generations: Antminer S19j Pro and S21 Pro.

What follows is not marketing data or lab simulations. It is based on PDU-level measurements collected directly from active production racks across our fleet.

Important Disclaimer

The use of custom third-party firmware such as Vnish is recommended only for out-of-warranty units that are no longer eligible for manufacturer-provided repairs.

Installing custom firmware can potentially void any remaining warranty coverage from Bitmain or other OEMs. Always confirm warranty status before modifying equipment.

The observations below reflect testing performed exclusively on out-of-warranty hardware and are shared for informational and educational purposes.

How Frontier Evaluates Firmware

When Frontier evaluates a firmware upgrade or tuning strategy, we apply the same process used for any operational optimization across our sites.

Power is captured every five minutes at the PDU or circuit level, paired with averaged hashrate data from miner telemetry. Efficiency is calculated using the standard formula:

Efficiency (J/TH) = Average Power (W) ÷ Average Hashrate (TH/s)

We remove outliers, stabilize profiles, and then deploy validated configurations across racks to observe real revenue impact over extended runtime.

This methodology allows us to move beyond anecdotal results and determine which configurations actually improve fleet economics at scale.

Throughout our ten years operating miners, we've run more than 200,000 ASIC miners running Vnish firmware across multiple sites and hardware generations. That scale of deployment gives our team a deep operational dataset for evaluating firmware behavior, stability, and performance across different power environments, cooling designs, and equipment lifecycles.

S19j Pro Results

Stock baseline: ~29.5 J/TH

Testing across multiple Vnish profiles produced the following results:

Key observations

• Efficiency peaks sharply around the 2000 W / ~80 TH range. 
• Beyond that point, incremental hashrate gains come with a clear efficiency penalty. This pattern is consistent with what we’ve observed across the broader S19 hardware family operating in our fleet. 
• Once tuned, these units have consistently delivered ~28.0 J/TH with strong stability and uptime in production environments. 

S21 Pro Results

Stock baseline: ~15.0 J/TH

Using Vnish firmware version 1.2.7 (2026 profiles), testing produced the following:

Key observations

• The S21 Pro already begins with strong stock efficiency, but firmware optimization can still deliver measurable improvements. 
• The 2910 W / ~206 TH profile produced the strongest efficiency results during our testing. 
• For miners that prefer to remain within the 220–245 TH performance band, the 3400–3650 W profiles still maintain excellent efficiency in the 14.4–14.7 J/TH range. 

Results Disclaimer

The performance profiles above reflect testing conducted under unique operating conditions specific to Frontier Mining's sites under management. Results may vary depending on a number of external factors, including:

• Facility design and airflow characteristics
• Cooling infrastructure and ambient climate
• Power distribution and electrical stability
• Fleet age and individual hardware condition
• Miner placement and rack density

While the relative performance trends are consistent, exact efficiency numbers and optimal profiles will differ across facilities and environments.

Stabilizing Consolidated Units with Vnish

Another area where we’ve seen Vnish deliver clear operational benefits is after miner consolidations.

Following consolidation, some hashboards may contain individual chips with slightly weaker performance characteristics. Under stock firmware, these boards often begin hashing normally but then trigger “unbalanced” errors after 15–30 minutes of runtime, causing instability or reduced output.

Vnish’s more granular autotuning and chip-level management help accommodate these variations, allowing consolidated units to maintain stable operation.

For miners performing repairs or fleet consolidation, this capability can help recover hashrate and extend the productive life of hardware that would otherwise remain unstable.

Why Firmware Strategy Matters Over a Miner’s Lifecycle

One of the most common mistakes we see in mining operations is treating hardware configuration as a one-time decision. In practice, the optimal operating profile for a miner evolves over time as hardware ages, network difficulty increases, and fleet composition changes.

The first stage we typically focus on is post-warranty performance tuning. After the warranty period ends, firmware adjustments can unlock performance and efficiency improvements that stock configurations often leave on the table, allowing operators to better align miner output with their site’s power strategy and infrastructure. 

• As fleets move further into their operating life, attention typically shifts toward efficiency optimization. Difficulty continues to climb, power strategies evolve, and operators increasingly prioritize improved J/TH over peak output. This stage is also where we tend to see the highest volume of repairs, hashboard replacements, and fleet consolidations. Over time, fleets become more heterogeneous as repaired boards and mixed components are reintroduced, making firmware with strong autotuning capabilities particularly valuable for maintaining stability across hardware with varying chip performance. 

Later in the lifecycle, the focus often turns to extending productive runtime. Rather than pushing for maximum performance, operators may choose more efficient, lower-power configurations that keep aging hardware economically viable. This stage frequently includes repaired or consolidated units, where firmware capable of accommodating chip-level variance can help stabilize boards that would otherwise struggle under stock settings. 

Ultimately, firmware tuning is not just about improving efficiency in the moment. It is about applying the right configuration at the right stage of a miner’s lifecycle — a strategy that helps fleets remain productive and competitive long after their initial deployment. 

What This Means for Operators

Across hardware generations, several consistent operational themes emerged from our testing.

• Moderate tuning strategies tend to yield the greatest efficiency gains.

  Chasing maximum hashrate almost always introduces steep efficiency penalties. In most environments, balanced profiles deliver a more sustainable combination of output, stability, and energy performance.

• Firmware flexibility becomes increasingly valuable as fleets age.

  As miners move beyond their warranty period and into mid-life operation, fleets often experience higher repair rates, hashboard replacements, and consolidations. Hardware inevitably becomes more varied, and firmware that can accommodate chip-level performance differences can help maintain stability across mixed-component units.

• Environmental factors matter more than many operators realize.

  Optimal profiles will vary depending on facility design, cooling architecture, climate, power infrastructure, and overall fleet condition. Results observed in one location should always be validated within the context of each specific site.

• Small efficiency gains compound across large fleets.

  Even modest improvements in J/TH can translate into meaningful operational advantages when deployed across thousands of miners operating continuously.

For operators managing large-scale deployments, the takeaway is simple: firmware strategy should evolve alongside the fleet itself. The most effective operations continuously reassess hardware performance and adjust tuning strategies as equipment moves through different stages of its lifecycle.

The Bigger Picture

Firmware tuning is just one piece of a larger operational puzzle.

At Frontier, our role as an operator is not only to keep machines hashing, but to continuously evaluate how fleets perform across their entire lifecycle — from initial deployment to long-term optimization.

The difference between a fleet that merely runs and a fleet that continues to produce strong economic returns years later often comes down to these kinds of operational decisions.

Curious how your fleet could perform with optimized firmware and power profiles? Our team regularly conducts fleet assessments to identify efficiency gains across different hardware generations. Reach out if you’d like us to take a look.