Skip to content

Releases: JuliaHealth/KomaMRI.jl

v0.9.0

01 Oct 18:09
ca742f6
Compare
Choose a tag to compare

KomaMRI v0.9.0

Diff since v0.8.2

This new breaking release of Koma v0.9 has multiple new features. Thanks to @rkierulf, our GSoC student, for the GPU enhancements and @pvillacorta for the new motion definitions that enable flow and shareable phantoms.

(1) Device-agnostic kernel-based simulations (more info here)
We can now use various GPU backends with the help of KernelAbstractions.jl. This means that GPU-related packages like CUDA are not installed by default. Installation and loading of the desired GPU package is needed (for example ] add CUDA, using CUDA).

(2) New simulation method, faster and better
Now Bloch has specialized implementations for CPU and GPU and the previous implementation is in BlochSimple. This new simulation method is 4-5 times faster and uses around 80 times less memory. We have seen a 100-200x improvement in some sequences, especially if they have a lot of soft RF pulses.

(3) Distributed simulations (more info here)
We also tested distributed simulations with multiple GPUs and nodes :)

(4) GPU benchmarks (here)
We have a fancy new benchmarking page to ensure that there are no performance regressions.

(5)Phantom's including flexible motions and flow (more experimental)
You can now plot phantoms with motion and define them using mix-and-match motion definitions! More complex phantoms are possible including simple motions or arbitrary motions. Many more examples of using them are coming soon! 😄 including a realistic CFD-simulated aorta! All of this is done using our new reproducible phantom definition (HDF5-based), which enables the sharing of these phantoms with complex motion.

Hope these changes make your simulations faster! 😄 As always, if you have any problems, feel free to talk to me (@cncastillo ) in Slack or post an issue or discussion on GitHub.

Merged pull requests:

Closed issues:

  • CUDA memory problems? (#125)
  • Using BuildKite for GPU related CI? (#147)
  • Use the same functions to sample RFs in the simulation than for plotting sequence (#167)
  • Separate CI for more comprehensive results (#276)
  • Add simulate function for debugging purposes (#288)
  • Create Pluto example for fat_sat_low_field (#291)
  • Reporting of phantom spatial sampling in obj.info structure. (#322)
  • GSOC Proposal Issue Tracking: Part 1 (#349)
  • GSOC Proposal Issue Tracking: Part 2 (#350)
  • Extend GPU support to include Apple (Metal.jl), AMD (AMDGPU.jl), and Intel (OneAPI.jl) (#351)
  • Profile Bloch Simulation Method (#352)
  • Create new Kernel-based Simulation Method (#353)
  • Use @LocalMem inside future kernel-based simulation functions to speed up memory access (#354)
  • Add example of Multi-GPU simulation (#355)
  • Add example and / or support for multi-node simulation (#356)
  • Simplify ArbitraryMotion struct (#371)
  • Re-use weights in interpolation for ArbitraryMotion (#372)
  • Define Nspins for SimpleMotion (#376)
  • Applying @time to simulate reveals discrepancies on resources and timing reports (#392)
  • plot_signal docstring (correction) (#401)
  • [KomaUI Plotting] obj_ui[] does not draw on first time updated. Need to press :rho or other option. (#407)
  • New motion approach to combine SimpleMotion and ArbitraryMotion into the same phantom (#410)
  • Future: use @testsetup to reduce time to run tests on GPU (#414)
  • Problems initializing KomaUI (#432)
  • Brain phantom values inconsistencies (#464)
  • spin_reset has eltype T and should be an AbstractMatrix{Bool} (#479)

v0.8.2

22 Apr 20:56
9befece
Compare
Choose a tag to compare

KomaMRI v0.8.2

Diff since v0.8.1

Merged pull requests:

Closed issues:

  • Modify Literate Examples Generation (#251)
  • plot_seq not plotting Gx if Gy is not on and vice-versa (#374)
  • KomaMRI errors when using Julia nightly (#375)
  • Simulation crashes when trying to simulate soft-RF (get_rf_center) (#378)

v0.8.1

12 Apr 17:10
bd2c310
Compare
Choose a tag to compare

KomaMRI v0.8.1

Diff since v0.8.0

Merged pull requests:

Closed issues:

  • Incorrect interpretation of time shaped RFs (#200)
  • Empty sequence adds an undesired block (#203)
  • Unnecessary samples are considered when concatenating sequences (#205)
  • Resolve Circular Dependencies in Tests Environments (#237)
  • Google Summer of Code (GSoC) (#243)
  • Refactor functions for computing moments (#268)
  • Compatible, tested or recommended julia versions for CPU and GPU? (#295)
  • Default simulate fails after clean install of KomaMRI version 0.7.5 (#298)
  • MRIReco.jl v0.8.0 incorrect initialization with multiple threads (#299)
  • Update codecov to consider new subpackages (#302)
  • Reduce spins for pelvis Phantom (#306)
  • Interpretation of '.seq' files is not matching with sequence designed on PyPulseq (#320)
  • Warning: sort(d::Dict; args...) is deprecated, use sort!(OrderedDict(d); args...) instead. (#326)
  • Can't define colormap ranges when plotting phantom object (#328)
  • [BUG] CodeCov Not Working (#335)
  • Refactor get_Mk to compute 't' and 'tadc' more efficiently (#345)
  • Performance improvements for plot_seq (function and UI) (#365)

v0.8.0

18 Jan 22:09
Compare
Choose a tag to compare

KomaMRI v0.8.0

Diff since v0.7.5

Merged pull requests:

Closed issues:

  • Annoying random update of plots in GUI (#66)
  • Separate KomaMRI into simpler subpackages (#92)
  • Add contributors to UI (#191)
  • Be more clear about the package versions in the UI (#192)
  • Koma logo in the docs is incorrect (#193)
  • Update pipeline figure to show the current organization of the Koma submodules (#194)
  • GUI gets stuck when simulating twice (#198)
  • Unattractive group of legends and colors in subplots (#199)
  • Add Contributing Guidelines (#209)
  • Add the ability to modify the seq, obj, sys, raw and image used in the UI from the REPL (#211)
  • MRD Data opened in Python appears corrupted (#214)
  • StackOverflowError on plots of large sequences (#221)
  • Precompiling KomaMRIPlots hangs on Windows (#225)
  • Add a template for reporting issues (#228)
  • KomaMRIFiles (#232)
  • Additional <userParameters> tag within XML header of exported MRD (#236)
  • Upload educational Pluto notebook (#242)
  • Update Subpackages in README (#244)
  • Update CompatHelper to also check subpacakge's compat (#245)
  • Add MRI together's talk to News section in README (#246)
  • Generete a test for slice simulation (#252)
  • Avoid Warnings During Document Generation (#255)
  • KomaMRIBase (#265)
  • Basic trapezoidal Grad usage, parameter definition with a plot. (#270)
  • KomaMRIPlots doesn't precompile in Julia 1.10.0 (#274)
  • Move Julia compat to 1.9 (#275)
  • Use nightly tests in CI (#281)
  • Increment code coverage (#282)
  • Improve the visualization of contributors and colors in the UI (#293)

v0.7.5

26 Oct 03:11
Compare
Choose a tag to compare

KomaMRI v0.7.5

Diff since v0.7.4

Merged pull requests:

Closed issues:

  • Edge cases arise during simple 1-spin simulation (#85)
  • Plots fail when used in Pluto notebooks (#166)
  • Default KomaUI sequence is different in julia versions 1.6 and 1.9 (#175)
  • Long file names overflow from UI's navbar (#176)
  • Adding code style, compatible Julia version, and license to README.md (#188)
  • Add KomaMRI.jl to MRHub (#196)
  • Bug with Δf in RF_sinc function (#197)
  • Replicate pulseq functions to create sequences (#206)

v0.7.4

13 Jul 11:59
c7c3f24
Compare
Choose a tag to compare

KomaMRI v0.7.4

Diff since v0.7.3

Closed issues:

  • Exporting results to .mat from the GUI (#70)
  • Solve edge-case for avoiding additional delay in the acquisition (#88)
  • Separate simulation base to KomaMRIBase subpackage (#92)
  • Problem with ADC next to RF object (#161)
  • Discretization of gradient waveform fails for square gradients / very high slew rates (#162)
  • KomaUI not working in Julia 1.9 (#168)

Merged pull requests:

  • Modules KomaMRICore and KomaMRIPlots (#148) (@beorostica)
  • Fix for no-gpu devices (#153) (@beorostica)
  • CompatHelper: add new compat entry for KomaMRICore at version 0.7, (keep existing compat) (#154) (@github-actions[bot])
  • Change UUID of KomaMRIPlots subpackage (#156) (@beorostica)
  • Add compat version 0.7 for KomaMRICore (#157) (@beorostica)
  • Apidocs (#158) (@beorostica)
  • CompatHelper: add new compat entry for KomaMRIPlots at version 0.7, (keep existing compat) (#159) (@github-actions[bot])
  • Update ci.yml for codecov subpackages (#160) (@beorostica)
  • KomaMRI v0.7.4: Bug fixes and minor features (#163) (@cncastillo)
  • Avoid dialog (#171) (@beorostica)
  • CompatHelper: add new compat entry for MAT at version 0.10, (keep existing compat) (#172) (@github-actions[bot])
  • Update Project.toml for KomaMRIPlots to v0.7.6 (#173) (@beorostica)

v0.7.3

05 Jan 18:36
Compare
Choose a tag to compare

KomaMRI v0.7.3

Diff since v0.7.2

Closed issues:

  • Two definitions for DiscreteSequence #146 (@cncastillo)
  • Hwloc still being used in KomaUI #144 (@cncastillo)
  • Blink v0.12.6 removed functions AtomShell.isinstalled() and AtomShell.install() used in KomaMRI.jl #143 (@cncastillo)

v0.7.2

30 Dec 19:01
Compare
Choose a tag to compare

KomaMRI v0.7.2

Diff since v0.7.1

Closed issues:

Merged pull requests:

v0.7.1

22 Dec 13:36
Compare
Choose a tag to compare

KomaMRI v0.7.1

Diff since v0.7.0

Closed issues:

  • DisplayFunctions: RF phase is incorrectly calculated (#130) @cncastillo
  • DisplayFunctions: Timings of RF waveform are slightly incorrect when rf_max_samples is less than the actual number of samples (#131) @cncastillo
  • Signal was being conjugated (affected "return_type" => "raw") (#129) @cncastillo
  • Pulseq read_seq prints unnecessary spaces (#133) @cncastillo
  • KomaUI prints some unnecessary spaces (#135) @cncastillo
  • Clean /examples/ folder (#137) @cncastillo

v0.7.0

08 Dec 16:42
8c513e7
Compare
Choose a tag to compare

KomaMRI v0.7.0

Diff since v0.6.7

Closed issues:

  • RF implementation with GPU acceleration (#2) @cncastillo
  • Documentation versions (stable & dev) (#115) @beorostica
  • Move gpu function calls to simulate (#116) @cncastillo
  • Solve edge-case for one acquisition sample (#117) @cncastillo
  • T2 decay was incorrectly calculated for the first sample after a delay. (#118) @cncastillo
  • Move get_grad and get_rfs function calls to simulate, to used them only once. (Low priority) (#119) @cncastillo
  • Type stability in core simulation types and functions (Phantom, DiscreteSequence, Magnetization). (#120) @cncastillo

Merged pull requests:

  • Update make.jl (#114) (@beorostica)
  • CompatHelper: add new compat entry for Adapt at version 3, (keep existing compat) (#122) (@github-actions[bot])
  • CompatHelper: add new compat entry for Functors at version 0.4, (keep existing compat) (#123) (@github-actions[bot])
  • CompatHelper: add new compat entry for Parameters at version 0.12, (keep existing compat) (#124) (@github-actions[bot])