This is a list of the material that has been produced regarding UEDGE performance.
The dates listed are the dates they were uploaded, and not necessarily the dates they were produced, although the distintion is unlikely to be significant.
If you would prefer more information be provided, or the explanations organized differently, lemme know.
I'm also thinking about cleaning this up some, so if you think there are files that we'll no longer need, they can be removed.

• compApr28.zip - 4/28/10 : Jacobian comparison results for 6-variable multistepping case with different comparison options activated.
• compMay13.zip - 5/13/10 : Jacobian comparison results for 5-variable multistepping case with 1,2,8 processors.
• compMay14.zip - 5/14/10 : Jacobian comparison results for 5-variable multistepping case with 1,2,8 processors. (Newer comparison version)
• compJun8.zip - 6/8/10 : Jacobian comparison results for mesh_64x32 case with 3,4,5 variables.
• compJun9.zip - 6/9/10 : Jacobian comparison results for mesh_64x32 case with 3,4,5 variables. (Newer comparison version)
• m64Jun14.zip - 6/14/10 : Initial statistics and SNESView output for 3 variable, mesh_62x32 case.
• m64picsJun15.zip - 6/15/10 : Graphs of 3 variable, mesh_64x32, LU solver run times for PETSc dense Jacobian and stencil Jacobian.
• m64picsJun28.zip - 6/28/10 : Graphs of 3 variable, mesh_64x32, ASM solver run times for PETSc dense Jacobian and stencil Jacobian.
• m64picsJul9.zip - 7/14/10 : Bar graphs and line graphs of mesh_64x32 with dense and stencial Jacobian and lag=5. Readme is included.
• m64picsJul21.zip - 7/21/10 : Redesigned bar graphs using the processor scaling so that not all bars are equal height.
• m128picsJul22.zip - 7/22/10 : Graphs of mesh_128x64 with stencil Jacobian, lag=5, LU, and 3-5 variables activated.
• m641DpicsJul23.zip - 7/23/10 : All mesh_64x32, lag=5 graphs redone with the 1D domain decomposition for NP=8.
• asmlu64.zip - 7/30/10 : Graphs for mesh_64x32, ASM with LU on blocks, 4 and 5 variables, lag of 1 and 3.
• m64lag1comparisons.zip - 8/1/10 : Comparing the LU and ASM+LU for mesh_64x32, 4 and 5 variables, lag=1. Also provides the KSP/SNES number.
• m128lag1.zip - 8/3/10 : Comparing the LU and ASM+LU for mesh_128x64, 3, 4 and 5 variables, lag=1. Argues that the neutrals are a real problem.
• m128smallstep.zip - 8/14/10 : LU and ASM comparison with a finite (10^-5) time step. Greatly improves ASM convergence.
• m128dt3-5.zip - 8/16/10 : LU and ASM comparison with dt=1e-3, 1e-4 and 1e-5. With neutrals off we see ASM performance.
• m128dt1-5.zip - 8/16/10 : LU and ASM comparison with all dts. All scalability and iteration values in an included excel file.
• neon10oct.zip - 10/10/10 : First set of tests with Neon. LU and FS(ASM on non-neutral/LU on neutral) comparison with dt=1e4 and 1e-4.
• dpptalk.pdf - 11/8/10 : This is the talk I gave at the APS-DPP conference in Chicago.
• siamcse11poster.pdf - 03/03/11 : This is the poster I presented at the SIAM CSE11 meeting in Reno.