BladeBit vs MadMax in 2026: Which Plotter Should You Use?

Key Takeaways:

• BladeBit offers three modes (RAM, Disk, CUDA) with compressed plot support up to C9, while MadMax creates uncompressed plots only using a fast CPU pipeline
• Choose BladeBit CUDA if you have an NVIDIA GPU with 8GB+ VRAM (CUDA 5.2+) and want compressed plots for maximum storage efficiency — plot times of 1–5 minutes
• Select MadMax for fast, reliable CPU-only plotting when you need simple uncompressed plots or your system lacks a compatible NVIDIA GPU
• Both plotters outperform the original ChiaPoS reference plotter by 10–50x, but serve different hardware setups and farming strategies
• Your existing hardware is the deciding factor: GPU systems favor BladeBit CUDA, while high-thread CPU rigs work well with MadMax
• Note: BladeBit CUDA is supported on Windows and Linux only — macOS is not currently supported for GPU plotting

Article Summary: BladeBit and MadMax represent two different approaches to Chia plotting in 2026. BladeBit offers flexible modes including GPU-accelerated compressed plotting (up to C9), while MadMax provides fast CPU-focused performance for uncompressed plots. Your choice depends on your hardware, whether you need plot compression, and your long-term farming goals.

Understanding Chia Plotting Software in 2026

The Chia plotting landscape has transformed dramatically since the blockchain’s 2021 launch. Early farmers struggled with the reference ChiaPoS plotter, which took hours to create a single k32 plot and wore down consumer SSDs quickly. Today’s plotting tools solve these problems through optimized code and smart hardware utilization.

Two plotters dominate the Chia ecosystem in 2026: BladeBit and MadMax. Both emerged from community development efforts and are now integrated into the official Chia software distribution. They offer plotting speeds far beyond the original reference plotter. However, they take fundamentally different approaches to creating plots, making each ideal for specific hardware configurations and farming strategies.

The Plotting Performance Revolution

The original ChiaPoS plotter could take 12 hours or more to create a single plot. The 2021 community plotting competition sparked innovation that produced MadMax — a plotter capable of completing a k32 plot in around 15 minutes on good hardware. Shortly after, Harold Brenes developed the original BladeBit as an in-memory plotter, pushing speeds even further on systems with substantial RAM. Chia Network Inc. eventually hired Brenes and expanded BladeBit into the multi-mode tool it is today.

By 2026, both plotters have matured into production-ready tools officially supported by Chia Network Inc. Farmers can now choose between multiple plotting approaches based on their specific hardware capabilities rather than being limited to a one-size-fits-all solution.

Why Plotter Choice Matters for Farmers

Your plotter selection impacts three critical farming factors: initial farm setup time, ongoing replotting capability for compression upgrades, and hardware wear from repeated plotting operations. A fast plotter reduces the weeks-long process of building a multi-petabyte farm to just days, while also enabling farmers to adapt their farms as new compression options become available.

Compressed plots save storage space but require more computing power during farming operations. This trade-off makes plotter choice strategic rather than purely technical. Farmers must balance upfront plotting speed against long-term operational costs and storage efficiency. If you want to dig deeper into which compression level suits your farm size, check out our guide on Chia compressed plots: trade-offs and settings for a detailed breakdown.

BladeBit: The Multi-Mode Chia Plotter

BladeBit stands out for its versatility, offering three distinct plotting modes that adapt to different hardware configurations.¹ Developed by Chia Network Inc. and maintained as an official plotting tool, BladeBit provides options for a wide range of farmer system setups — from RAM-loaded servers to gaming desktops with NVIDIA GPUs.

BladeBit RAM Mode

BladeBit’s RAM mode creates plots entirely in system memory, eliminating disk writes and achieving very fast plotting speeds. This mode requires 416 GB of available RAM — placing it firmly in workstation and server territory.¹ Plotting happens entirely in memory at speeds typically ranging from 2–5 minutes per k32 plot, limited only by CPU performance.

RAM mode supports compressed plots up to C9, allowing farmers to create space-efficient plots on high-memory systems with no GPU required. The mode suits dedicated plotting workstations or servers but remains impractical for typical desktop farming setups due to the substantial memory requirements.

BladeBit Disk Mode

Disk mode brings BladeBit’s performance to mainstream systems with as little as 2 GB of RAM (though higher RAM allows larger cache allocations that improve speed). This mode uses temporary SSD or HDD storage for plotting operations while incorporating a DRAM write cache that significantly reduces disk writes.¹ The architecture employs sequential write patterns that maximize SSD endurance and reduce write amplification.

Disk mode requires approximately 480 GB of temporary disk space in default mode, or 390 GB with the --alternate flag. It produces uncompressed plots in Chia 2.0, with compressed plot support introduced in Chia 2.1. In practice, most farmers wanting compressed plots use BladeBit CUDA rather than Disk mode, as CUDA is significantly faster for that purpose.

BladeBit CUDA Mode

CUDA mode leverages NVIDIA GPU acceleration to create compressed plots at rapid speeds. It requires an NVIDIA GPU with CUDA capability 5.2 or higher and at least 8 GB of VRAM — tested GPUs in the official documentation include the GTX 1060, RTX 2060, RTX 3060 Ti, RTX 3090, and A4000, among others.⁶ Plot times range from approximately 1–5 minutes on a supported GPU with full 256 GB RAM, or 3–5 minutes in the RAM + SSD hybrid configuration using 128 GB of RAM.

CUDA mode supports compression levels up to C9, which is the highest available in the official Chia plotter ecosystem.¹ An experimental low-memory option using as little as 16 GB RAM + SSD is available in the standalone BladeBit build, though this is not officially supported in the same way as the 128 GB and 256 GB configurations. Important note: BladeBit CUDA is supported on Windows and Linux only — macOS is not currently supported for CUDA plotting.

MadMax: The CPU Performance Plotter

MadMax revolutionized Chia plotting in 2021 by introducing a pipeline architecture that maximizes CPU thread utilization across all plotting phases.⁵ This design allows the plotter to achieve speeds far beyond the original ChiaPoS reference plotter on modern multi-core processors, with real-world k32 plot times ranging from 15–60 minutes depending on hardware configuration.

The plotter requires minimal configuration and works across Windows, macOS, and Linux systems. Unlike BladeBit’s multiple modes, MadMax follows a single straightforward approach: use as many CPU threads as possible to process plot data through an optimized pipeline. It creates uncompressed plots only.

MadMax Architecture and Performance

MadMax operates using two temporary directories: tmpdir handles the first phase of writes while tmpdir2 manages the majority of the temporary data. This split allows farmers to use a RAM disk for the larger tmpdir2, dramatically reducing SSD writes and extending drive lifespan.⁵ Total temporary space requirements sit around 330 GB across both directories.

Plot times vary based on CPU thread count and memory configuration. The official Chia docs place CPU + SSD plotting in the 15–60 minute range, with optimal multi-thread setups hitting the lower end of that range. The plotter scales reasonably well with thread count, making it a solid choice for AMD Threadripper and Intel Xeon systems with high core counts.

RAM Disk Optimization

MadMax’s efficiency advantage comes largely from RAM disk usage. Allocating substantial system RAM as a tmpdir2 RAM disk significantly reduces SSD writes, extending consumer SSD lifespan. This makes MadMax a practical choice for farmers with 128 GB or more RAM but no compatible NVIDIA GPU for BladeBit CUDA.

Setting up a RAM disk is straightforward on both Linux and Windows. Once configured, MadMax automatically uses the RAM disk for its write-intensive operations, achieving better plotting speeds while protecting consumer-grade SSDs from premature wear.

MadMax Limitations in 2026

MadMax does not support compressed plot creation — this is its primary limitation in the current Chia landscape. Farmers wanting compressed plots must choose BladeBit CUDA, BladeBit RAM, or a third-party plotter like Gigahorse (which carries a dev fee) or DrPlotter (officially supported by Chia Network). This restriction means MadMax serves the uncompressed farming segment, which remains a valid choice for farmers who value simplicity and operational transparency over maximum storage efficiency.

The plotter also shows diminishing performance returns beyond 32–40 active threads in most setups. While it scales reasonably with core count, improvements become marginal beyond that range, making parallel plotting with multiple instances a consideration for extreme high-core-count systems.

Hardware Requirements Comparison

Understanding hardware requirements helps farmers select the right plotter for their existing equipment. Both BladeBit and MadMax create k32 plots, but their resource needs differ substantially across modes. For a deeper look at how different storage media affect plotting pipelines, see our NVMe vs HDD plotting pipeline guide.

Memory Requirements

BladeBit RAM mode requires 416 GB — placing it firmly in workstation and server territory.¹ BladeBit CUDA mode requires either 256 GB for full in-memory GPU plotting, or 128 GB for the supported hybrid disk configuration. An experimental 16 GB option exists in the standalone build but is not fully supported. BladeBit Disk mode requires a minimum of just 2 GB, though more RAM allocated as cache improves performance significantly.

MadMax operates with very minimal RAM at its base — technically as little as 0.5 GB in single-threaded mode per the official docs. However, it delivers its best real-world performance when 128 GB or more is available to use as a RAM disk for the tmpdir2 directory, protecting SSDs from heavy write cycles.

GPU and CPU Considerations

BladeBit CUDA requires an NVIDIA GPU with CUDA capability 5.2 or higher and at least 8 GB of VRAM.⁶ Officially tested GPUs include the GTX 1060, RTX 2060, RTX 3060 Ti, RTX 3090, RTX 4060, Tesla P4, and A4000. AMD GPUs are not currently supported for CUDA plotting. BladeBit CUDA runs on Windows and Linux only — macOS is not supported.

MadMax relies purely on CPU performance and scales with thread count. A modern 8-core CPU produces decent results, but 16–32 threads deliver the sweet spot for speed versus cost. Intel Xeon and AMD Threadripper processors excel with MadMax, as do Ryzen 5000 and 7000 series chips with high core counts. macOS is supported for MadMax, unlike BladeBit CUDA.

Storage Requirements

Both plotters need temporary storage unless running fully in RAM. BladeBit Disk mode requires approximately 480 GB in default mode or 390 GB with the alternate flag. MadMax needs around 330 GB total across both temp directories, with the split depending on your RAM disk configuration.

Enterprise or data center NVMe SSDs provide better endurance than consumer drives for heavy plotting. BladeBit Disk’s sequential write pattern with low write amplification means even consumer NVMe drives can handle a meaningful number of plots before wearing out.¹ MadMax with a RAM disk similarly protects consumer SSDs by keeping the most write-intensive operations in memory.

Compressed Plots: The BladeBit Advantage

Plot compression is one of the most significant advances in Chia farming efficiency. Compressed plots store the same farming power in less physical space, allowing farmers to increase effective farm size without adding drives. However, compression requires more computing resources during harvesting.

BladeBit CUDA and RAM modes support compression levels C1 through C9, progressively reducing plot file size.¹ Higher compression levels yield smaller plots but demand more from your harvester during proof lookup. C9 plots are the most space-efficient at 75.2 GiB, compared to 101.4 GiB for uncompressed plots. MadMax cannot create any compressed plots.

Compression Level Trade-offs

Lower compression levels (C1–C3) provide moderate space savings with minimal farming overhead. These levels work well on standard farming hardware and reduce storage costs without requiring powerful harvesting systems. Mid-level compression (C4–C5) offers better space efficiency but demands more robust farming hardware, typically requiring GPU-assisted harvesting for larger farms.

Higher compression (C6–C9) maximizes storage efficiency but requires substantial harvesting resources. Farmers using C7–C9 plots typically deploy GPU harvesters to handle the decompression workload within proof time windows. The farming overhead versus storage savings calculation depends heavily on electricity costs, current XCH price, and available hardware.

When Compression Makes Sense

Farmers with high storage costs but low electricity costs benefit most from compressed plots. The upfront investment in GPU plotting hardware and ongoing harvesting power consumption pays off through reduced storage expansion needs. Conversely, farmers with cheap storage space but expensive electricity may prefer uncompressed plots and MadMax plotting.

Small to medium farms (under 1 PiB) can often run moderate compression levels effectively on a single GPU harvester. Large farms require careful planning to ensure harvesting infrastructure can handle the decompression workload without missing proof deadlines. Some farmers run mixed farms with both compressed and uncompressed plots to balance efficiency with operational simplicity.

Performance Benchmarks and Real-World Results

Understanding hardware requirements helps, but real-world performance varies based on specific configurations. Community benchmarks from 2023–2026 provide insight into what farmers actually achieve with different plotter and hardware combinations.

BladeBit CUDA Performance

Official Chia docs show that a GPU + RAM setup (256 GB) produces k32 plots in approximately 1–3 minutes, while a GPU + SSD hybrid (128 GB RAM) achieves 3–5 minutes.⁶ A Dell Precision 5860 workstation with Xeon W5-2455X CPU, 256 GB RAM, and RTX 3070 GPU achieved compressed plot times in the sub-3-minute range across all compression levels in independent testing.⁹ Budget builds around the HPE Z440 (256 GB RAM, RTX 2070) show estimated plot times of approximately 170–200 seconds per the official Chia hardware page.

Plot time increases with compression level — creating a C9 plot takes longer than a C1 plot on the same GPU, as more computation is involved. Most farmers standardize on mid-to-high compression levels that balance space savings with plotting throughput for production farming operations.

MadMax Performance Metrics

MadMax performance varies significantly based on CPU thread count and memory configuration. The official Chia hardware table shows CPU + SSD plotting in the 15–60 minute range.⁶ Systems with higher thread counts (16–32 threads) using a RAM disk for tmpdir2 hit the lower end of that range, while 8–12 thread systems with slower storage see longer times. RAM disk configurations provide meaningful improvements for both speed and SSD longevity.

Head-to-Head Comparison

For uncompressed plotting, BladeBit Disk and MadMax are competitive on the same CPU + SSD hardware. Community benchmarks on Threadripper Pro systems have shown BladeBit Disk in the 14–18 minute range, with MadMax producing competitive times when properly tuned with RAM disk.⁸ Individual results vary based on tuning, CPU generation, and storage throughput.

For compressed plotting, BladeBit CUDA has no CPU-only competition — MadMax simply does not support it. Farmers needing compressed plots choose BladeBit CUDA (official, no dev fee), DrPlotter (officially supported, GPU required), or Gigahorse (third-party, carries a dev fee).¹

Expert Perspective on Plotter Selection

Jonmichael Hands, who served as VP of Storage Business Development at Chia Network and co-authored the BladeBit 2.0 announcement, explained the design philosophy clearly: “We wanted to create a plotter that is high-performance, but had low minimum requirements and compatibility across a broad set of system configurations. We have a lot of Chia farmers using a mainstream desktop for plotting.”²

This philosophy drove BladeBit’s development of multiple modes serving different farmer segments. The official Chia team recognized that no single plotter could serve all farmers optimally, leading to support for BladeBit, MadMax, and DrPlotter in the official Chia software distribution.

Community Development Impact

The Chia plotting landscape demonstrates the value of community-driven development in blockchain ecosystems. MadMax emerged from an independent developer (madMAx43v3r) responding to farmer needs, while BladeBit originated from Harold Brenes’s innovation and was later adopted as an official Chia Network project. Both plotters continue receiving active development and optimization based on real farmer feedback.

Third-party tools like Gigahorse and DrPlotter further expand plotting options.¹ DrPlotter is officially listed in Chia documentation and supported by Chia Network Inc., while Gigahorse carries a dev fee but supports a wide range of k-sizes (k30–k34). The diversity of plotting tools ensures farmers can match their hardware and economic constraints to an optimal plotting strategy.

Making Your Decision: Which Plotter to Choose

Selecting between BladeBit and MadMax depends on your hardware configuration, farming strategy, and whether you need compressed plots. Most farmers can identify their optimal plotter by working through their system capabilities and goals.

Choose BladeBit CUDA If You Have

An NVIDIA GPU with 8 GB or more VRAM and a Windows or Linux system. BladeBit CUDA creates compressed plots (up to C9) far faster than any CPU-only option, making it the clear choice for farmers with a supported graphics card. Even gamers with RTX 3060 Ti or higher cards can plot effectively during non-gaming hours, building substantial farms without dedicated plotting hardware. Systems with 128 GB RAM can use the supported hybrid disk mode, while 256 GB RAM enables fully in-memory GPU plotting for the fastest speeds.

Choose BladeBit RAM If You Have

A workstation or server with 416 GB or more RAM but no compatible NVIDIA GPU. RAM mode provides fast in-memory plotting and supports compression up to C9, with no GPU required. This mode suits plotting-as-a-service setups or farmers building very large farms quickly who want compressed plots without GPU hardware.

Choose MadMax If You Have

A high thread-count CPU (16+ threads) with a solid RAM configuration but no suitable NVIDIA GPU. MadMax excels on AMD Threadripper and Intel Xeon systems, maximizing CPU utilization across all plotting phases. MadMax also supports macOS, which BladeBit CUDA does not. The RAM disk optimization makes it the most efficient CPU-only plotter for uncompressed plots. Farmers content with uncompressed farming avoid the complexity and ongoing power consumption of GPU harvesting while maintaining solid plotting speeds.

Choose BladeBit Disk If You Have

Limited system resources (2–32 GB RAM) and no GPU option. BladeBit Disk brings accessible plotting to entry-level systems, making Chia farming practical for newcomers. Its sequential write pattern protects consumer-grade SSDs better than many alternatives. This mode works well for casual farmers building small to medium farms on gaming desktops or general-purpose computers.

Conclusion: Both Plotters Serve Essential Roles

BladeBit and MadMax are complementary tools in the 2026 Chia plotting ecosystem rather than competing alternatives. Your hardware capabilities and farming strategy naturally point to the right choice. Farmers with compatible NVIDIA GPUs should leverage BladeBit CUDA for compressed plotting speed and storage efficiency. High-thread CPU systems without GPUs — or farmers on macOS — excel with MadMax’s efficient pipeline for uncompressed plots. If you have 416 GB of RAM and no GPU, BladeBit RAM gives you compressed plot capability through pure CPU power.

Take time to evaluate your specific hardware, consider whether compressed or uncompressed farming fits your long-term strategy, and test your chosen plotter on a small batch before committing to a full replot. Both tools are free, open-source, and officially supported — the best plotter is simply the one that fits your farm.

BladeBit vs MadMax 2026 FAQs

Can I use both BladeBit and MadMax on the same system?

Yes, you can use both BladeBit and MadMax on the same system, as both are integrated into the official Chia software. Many farmers run different plotters across different machines within their farm, or switch between them depending on whether they are creating compressed or uncompressed plots for different storage pools.

Which plotter is better for BladeBit vs MadMax in 2026?

In the BladeBit vs MadMax 2026 comparison, the best choice depends on your hardware: choose BladeBit CUDA if you have a compatible NVIDIA GPU (CUDA 5.2+, 8 GB VRAM) on Windows or Linux for fast compressed plotting in 1–5 minutes. Choose MadMax for CPU-only systems with 16+ threads for reliable uncompressed plotting in 15–60 minutes, and it also supports macOS unlike BladeBit CUDA.

Does BladeBit support higher compression than C7?

Yes — BladeBit CUDA and BladeBit RAM both support compressed plots up to C9, which is the highest compression level available in the official Chia plotter. C9 plots measure 75.2 GiB versus 101.4 GiB for uncompressed (C0) plots, but require the most compute resources during harvesting.

What RAM do I need for BladeBit vs MadMax?

BladeBit requirements vary by mode: 2 GB minimum for Disk mode, 128 GB or 256 GB for CUDA hybrid or full-RAM GPU plotting, and 416 GB for BladeBit RAM mode. MadMax can technically run on very low RAM but performs best when you allocate 128 GB or more as a RAM disk for the tmpdir2 directory, which significantly reduces SSD writes during plotting.

Does BladeBit CUDA work on macOS?

No — BladeBit CUDA is currently supported on Windows and Linux only; macOS is not supported for GPU plotting. BladeBit Disk and BladeBit RAM modes do support macOS, as does MadMax. If you are on macOS and want compressed plots, you would need to run BladeBit RAM mode on a system with 416 GB of RAM.

BladeBit vs MadMax 2026 Citations

1. Chia Network. “Plotting Software.” Chia Documentation. https://docs.chia.net/reference-client/plotting/plotting-software/
2. Hands, Jonmichael and Brenes, Harold. “Announcing Bladebit 2.0.” Chia Network Blog, August 8, 2022. https://www.chia.net/2022/08/08/announcing-bladebit-2-0/
3. Chia Network. “BladeBit GitHub Repository.” GitHub. https://github.com/Chia-Network/bladebit
4. XCH.farm. “Chia Plotting Overview.” XCH.farm Resources. https://xch.farm/plotting/
5. madMAx43v3r. “Chia Plotter GitHub Repository.” GitHub. https://github.com/madMAx43v3r/chia-plotter
6. Chia Network. “Plotting Hardware.” Chia Documentation. https://docs.chia.net/reference-client/plotting/plotting-hardware/
7. Chia Decentral. “GPU Plotting Build Guide.” Chia Decentral Resources. https://chiadecentral.com/gpu-plotting-build-guide/
8. GitHub Discussion. “New BladeBit 2.0 vs MadMax+RamCache.” BladeBit Discussions. https://github.com/Chia-Network/bladebit/discussions/186
9. Science of Mining. “BladeBit CUDA Performance Analysis.” Performance Testing, August 28, 2023. https://www.scienceofmining.com/20230828-BladeBit/BladeBit%20CUDA%20Performance%20Analysis-1.html
10. The Techtellectual. “Chia vs. Mad Max: Choosing the best Chia plotter for Linux.” Linux Guides, April 5, 2024. https://techtellectual.com/traditional-vs-mad-max-choosing-the-best-chia-plotter-for-linux/
11. XCH.farm. “Chia GPU Plotting.” GPU Resources. https://xch.farm/gpu-plotting/