diff --git a/README.md b/README.md index 1a76e0d9..303e1772 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ # Switch OC Suite -[![License: GPL v2](https://img.shields.io/badge/License-GPL_v2-blue.svg)](https://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html) [![Join the chat at https://gitter.im/Switch-OC-Suite/community](https://badges.gitter.im/Switch-OC-Suite/community.svg)](https://gitter.im/Switch-OC-Suite/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge) +[![License: GPL v2](https://img.shields.io/badge/License-GPL_v2-blue.svg)](https://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html) Overclocking suite for Horizon OS (HOS) running on Atmosphere CFW. @@ -13,22 +13,6 @@ This project will not be actively maintained or regularly updated along with Atm - Higher RAM clocks without proper timings could be UNSTABLE and cause graphical glitches / instabilities / filesystem corruption. **Always make backup before usage.** -- Why no CPU/GPU OC for Erista? -
- - - Tegra X1 on Erista is on TSMC 20nm HPM node, consumes much more power (~2x) and generates much more heat, compared to Tegra X1+ on Mariko (TSMC 16nm FinFET). - - Erista Switch uses lower speedo (=== lower quality === higher voltage required) SoC from NVIDIA. You will NOT get comparable performance to NVIDIA Shield TV no matter what. - - Snapdragon 810 (4 x A57 @ 2.0GHz + 4 x A53) also uses 20nm HPM, see how it plagued Android phones in 2014. - - - The board power supply is quite limited, even if you've done cooling mod. - - You could spot battery draining at higher clocks under stress test, even with official 39W PD charger. - - CPU / GPU performance at max clocks will be worse if power supply is not enough. - - - CPU OC (up to ~ 2.1 GHz, depending on your CPU bin) is available mainly for emulation, but it does NOT work out of the box. - -
- - ## Features - **DRAM Overclock** @@ -53,43 +37,17 @@ This project will not be actively maintained or regularly updated along with Atm - **[System Settings (Optional)](https://github.com/KazushiMe/Switch-OC-Suite/blob/master/system_settings.md)** -- **TinyMemBenchNX**: DRAM throughput and latency test based on [tinymembench](https://github.com/ssvb/tinymembench) - -- **MemTesterNX**: A userspace utility for testing DRAM faults and stability based on [memtester](https://pyropus.ca/software/memtester/) - - Now with multi-thread support and "stress DRAM" option, it should be able to test DRAM stability with adjusted timings. - ### Mariko Only - **CPU/GPU Overclock** - Safe: CPU/GPU @ 1963/921 MHz - - It has been proved safe without charger (not reaching battery power draw threshold) - - Unsafe: CPU/GPU up to 2397/1305 MHz + - Unsafe: **NOT RECOMMENDED** - **Disabled by default**, toggle "Allow Unsafe Frequencies" on in overlay or add `allow_unsafe_freq=1` to `config.ini` -
- - - Without chargers, CPU/GPU would be capped @ 1963/921 MHz or -/460 MHz (Emulator mode). - - - Without official chargers, GPU would be capped @ 1267 MHz. - - - Why **NOT RECOMMENDED**? - - See `Current Flow` in sys-clk-OC overlay `Miscellaneous` (on battery) or measure power draw from charger yourself. - - Currently, there are no protective measures from heavy power draw. - - - CPU: 2397 MHz @ 1220 mV (overvolting from 1120 mV) - - NVIDIA Official Maximum: 1963.5 MHz - - Maximum performance depends on CPU speedo. - - You'd get somewhere between 2360 to 2390 MHz performance for real. - - This is where floating point performance maxed out. - - ≥ 2193 MHz will _ENABLE OVERVOLTING_. - - - GPU: 1305 MHz (no overvolting, less than official threshold 1050 mV) - - NVIDIA Official Maximum: 1267.2 MHz - - ~~Tested with deko3d compute shaders converted from Maxwell SASS assembly. Single-precision floating point (FP32 FFMA) performance maxes out at 1305 MHz.~~ - -
+ - Power draw will be significant higher than what the mainboard was designed to tolerate at anything higher than 1963/921 MHz. + - See [README for sys-clk-OC](https://github.com/KazushiMe/Switch-OC-Suite/blob/master/Source/sys-clk-OC/README.md) - **Modded sys-clk and ReverseNX**(-RT)