maki/openairinterface5g
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
4,447 Commits 5 Branches 0 Tags
f3e084e005fd3bc7562ec5aabdc2adaa3be2c037
Commit Graph
706 Commits
Author SHA1 Message Date
Raymond Knopp 8d4405bddf added avx2 optimized turbo decoder for 16-bit LLR. This decoder parallelizes by decoding 2 code segments concurrently. requires updates dlsch_decoding.c to identify when new parallel version can be used. other minor changes related to memory allocations for future avx2 optimizations (32-byte alignment). 2016-01-03 15:15:26 -08:00
Raymond Knopp 27b1707ed1 AVX2 updates to turbo encoder interleaver. Also speedup in RSC encoder. 40% speedup improvement. 2016-01-01 05:51:09 -08:00
Rohit Gupta 8e1b5d26d4 doxygen support for USRP PHY Interface 2015-12-31 09:06:10 +01:00
Raymond Knopp 7f53f84c00 Tested modifications for 23.04 Ms/s 100 PRB transmission/reception (Downlink). This will reduce I/Q acquisition times for 20 MHz channels. new option in dlsim (-E) for testing 1536-point DFT instead of 2048 for this lower sampling rate. Added normalization to idft1536. Should also be added to dft1536 later. 2015-12-29 20:34:47 +01:00
Raymond Knopp 7f2dd71e68 Merge remote-tracking branch 'origin' into enhancement-57-UHD-interfacing 2015-12-29 20:29:47 +01:00
Raymond Knopp 93bf163071 commit prior to merge with develop 2015-12-29 20:29:26 +01:00
Raymond Knopp 1ff2c0d3f4 Merge branch 'develop' into feature-38-bladerf 
Conflicts:
	targets/RT/USER/lte-softmodem.c
 2015-12-27 11:42:41 +01:00
Raymond Knopp 596b0fd1c5 Merge branch 'develop', remote-tracking branch 'origin' into feature-38-bladerf 2015-12-25 09:57:51 +01:00
Cedric Roux 3d0a6e4032 fix a bug in dot_product. The return was done before _mm_empty() 2015-12-22 13:28:05 +01:00
gauthier cf8645d8cd Limitation of non SI DLSCH allocations in subframe 0/5 to PRBS not around DC (avoid puncturing from PBCH). Placement of SI around DC instead of randomized allocations. Placement of RA allocations at PRB 0-3. Tested DL throughput on 5 MHz to 16 Mbit/s in TM1. 2015-12-21 19:05:28 +01:00
gauthier a59cabd1dd Merge branch 'develop' into bugfix-48-L1L2signaling 
Conflicts:
	openair1/SCHED/phy_procedures_lte_eNb.c
	targets/PROJECTS/GENERIC-LTE-EPC/CONF/enb.band7.tm1.usrpb210.conf
 2015-12-21 18:45:27 +01:00
Florian Kaltenberger 6f889b30e4 minor changes to be able to load full frame 2015-12-18 17:26:48 +01:00
Florian Kaltenberger e5bcff484c first version of phy only test 2015-12-18 14:52:45 +01:00
Florian Kaltenberger 9dd329d842 Merge branch 'hotfix-58-bfly3bug' into 'develop' 
Applied patch to lte_dfts.c (inline function cmacc for x86) to correct bug in ibfly3



See merge request !7
 2015-12-17 14:52:31 +01:00
gauthier b0a54d8c7e fix of PUCCH statistic tracking with signal scope 2015-12-17 09:44:38 +01:00
gauthier f5a7184185 Merge branch 'bugfix-48-L1L2signaling' of https://gitlab.eurecom.fr/oai/openairinterface5g into bugfix-48-L1L2signaling 
Conflicts:
	openair1/PHY/LTE_TRANSPORT/proto.h
 2015-12-16 17:26:12 +01:00
gauthier ede2152862 Merge branch 'develop' into bugfix-48-L1L2signaling 
Conflicts:
	targets/ARCH/COMMON/common_lib.c
	targets/RT/USER/lte-softmodem.c
 2015-12-16 11:12:28 +01:00
Rohit Gupta 40f242094f modified due to compilation error 2015-12-14 16:38:10 +01:00
Raymond Knopp 1e08469ed8 conversion of some sse4 code to avx2 for turbo encoder. No noticeable improvement (output expansion for encoder doesn't benefit much because of unpack instruction difference on AVX2. 
maxh adjustement in dlsch_demodulation (to be checked further for other mcs)

exit criterion in dlsim to allow estimation of lower BLER.
 2015-12-14 01:51:26 -08:00
Raymond Knopp 5dbe3bef51 debugging and testing of AVX2-optimized DFTS (radix 2/4 only). 2015-12-13 07:57:04 -08:00
Cedric Roux 7290d1bc89 Merge branch 'develop' into bugfix-56-UE_category3_throughput 2015-12-07 12:08:20 +01:00
Rohit Gupta ef498d78b1 Merge branch 'develop' of https://gitlab.eurecom.fr/oai/openairinterface5g into develop 2015-12-06 23:12:26 +01:00
Raymond Knopp 47a6fa6efd Merge branch 'develop' into enhancement-43-AVX2 2015-12-05 02:44:33 -08:00
Raymond Knopp cb0eabe7df updated unitary simulations for new CCE interfaces. Cleaned up warnings in dlsim/pdcchsim/ulsim. 2015-12-05 02:28:21 -08:00
Cedric Roux a31432e37e some cleanup 2015-12-04 09:15:08 +01:00
Cedric Roux f630fc5863 Merge branch 'develop' into bugfix-56-UE_category3_throughput 2015-12-02 08:38:10 +01:00
Florian Kaltenberger 9eeff7210d Merge branch 'bugfix-40-channel_bandwidth' into develop 2015-11-30 17:00:09 +01:00
Raymond Knopp 11cfa5ba76 brought in AVX2 modifications from svn repository on openair1/PHY/TOOLS/lte-dfts.c 2015-11-29 13:08:08 -08:00
Raymond Knopp 1fc396f57d Merge after bad rebase 
Merge branch 'bugfix-48-L1L2signaling' of https://gitlab.eurecom.fr/oai/openairinterface5g into bugfix-48-L1L2signaling

Conflicts:
	openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation.c
	openair1/SCHED/phy_procedures_lte_ue.c
	openair2/LAYER2/MAC/eNB_scheduler.c
	openair2/LAYER2/MAC/eNB_scheduler_RA.c
	openair2/LAYER2/MAC/eNB_scheduler_bch.c
	openair2/LAYER2/MAC/eNB_scheduler_dlsch.c
	openair2/LAYER2/MAC/eNB_scheduler_primitives.c
	openair2/LAYER2/MAC/eNB_scheduler_ulsch.c
	openair2/LAYER2/MAC/proto.h
	openair2/PHY_INTERFACE/defs.h
 2015-11-29 01:06:14 -08:00
Raymond Knopp f48d10ba3e Added DLSCH scheduling for subframes 0 and 5. Changes in MAC scheduling data structure to keep track of VRB allocations from SI,RA when scheduling regular DLSCH 2015-11-29 00:37:04 -08:00
Raymond Knopp f7c67795ff CCE Allocation in MAC, rebase from develop. Frame number limitation to 1024 for non-real time targets (caused bad bug in oaisim). 2015-11-25 13:49:40 -08:00
Raymond Knopp b10937db82 Moved CCE allocations from PHY to MAC and updated MAC interfaces for BCCH/RA/UE-SPEC scheduling. 2015-11-24 14:35:20 -08:00
Raymond Knopp e3030b0f9f corrected HARQ ACK/NAK handling for TDD, in openair1/SCHED/phy_procedures_lte_common.c. 2015-11-24 14:35:20 -08:00
ABEILLE 4f35207900 merged in changes from abeille (modifications on final svn version) 2015-11-24 14:35:20 -08:00
Raymond Knopp ccef9862fa nCCE changes prior to rebase 2015-11-24 14:34:32 -08:00
Cedric Roux bae7f1ed2c Merge branch 'develop' into bugfix-56-UE_category3_throughput 2015-11-23 16:56:06 +01:00
Florian Kaltenberger 3d21f622f7 Merge branch 'develop' into bugfix-40-channel_bandwidth 
Conflicts:
	openair1/PHY/LTE_TRANSPORT/proto.h
 2015-11-23 11:58:44 +01:00
Cedric Roux f14a401f27 Manage free downlink HARQ processes with a circular list. 
The oldest PID (the PID that was used the longest
TTI ago) is at the head and is the next one that
will be used.

No test is done to check that when a process is
allocated it is at least N TTIs old, but some quick
thinking strongly leads to the 'conclusion' that N
won't be smaller than 8, which seems to be what
cat3 UEs require to function properly in the downlink
direction (from the eNodeB to the UE).
 2015-11-23 11:57:03 +01:00
Rohit Gupta 4b7dda33c2 Merge branch 'feature-34-test_framework' into develop 2015-11-23 11:18:53 +01:00
Rohit Gupta 1c0a59ffcd Merge branch 'develop' into feature-34-test_framework 2015-11-23 11:18:13 +01:00
aikaterini.trilyraki 9db853f1b1 changes to lte-softmodem proposed by alcatel- minor change to defs.h 2015-11-19 20:15:40 +01:00
Raymond Knopp 75b32aad87 Applied patch to lte_dfts.c (inline function cmacc for x86) to correct bug in ibfly3 2015-11-19 00:32:18 -08:00
Raymond Knopp 2fe218e91a Moved CCE allocations from PHY to MAC and updated MAC interfaces for BCCH/RA/UE-SPEC scheduling. 2015-11-18 14:18:58 -08:00
Navid Nikaein 23bb81d2d0 change the printf to LOG_D for the rssi message 2015-11-18 15:48:31 +01:00
Cedric Roux 68249569ef Give a better name to this array. 2015-11-10 10:34:16 +01:00
Cedric Roux f2b3597b55 This is a very basic fix, more work is needed. 
The problem seems that we reuse HARQ processes too early.
For example, at subframe 0 we allocate PID 0. At subframe 4
we receive an ACK. At subframe 5 we may well reallocate this
PID. It seems category 3 UEs don't like that. They expect
some delay. How much? I don't know. 8 maybe, as for UL.

This commit forces allocation of HARQ PID:
  0 on subframe 1,
  1 on subframe 2,
  2 on subframe 3,
  3 on subframe 4,
  4 on subframe 6,
  5 on subframe 7,
  6 on subframe 8,
  7 on subframe 9.

We don't use subframes 0 and 5 (for initial transmission at
least). (Current develop branch doesn't either I think.)

This is not a good solution, just a quick and dirty one. With this
commit I can achieve 12Mbps with iperf UDP on a 5MHz band 7 carrier
with a cat3 UE. And more than 11Mbps with iperf TCP. And a bad
radio link.

We may want to implement some sort of free-list and take the
oldest PID in there, if it is older than let's say 8 subframes
(that is: the last transmission with this pid was done more
than 8 subframes ealier). We may well have no free PID if a
lot of retransmissions are done.
 2015-11-10 10:19:07 +01:00
Florian Kaltenberger 916008f47d fixing a few compilation warnings in ulsim 2015-10-28 11:28:25 +01:00
Florian Kaltenberger a725a3a86b some error checks in ulsim added 2015-10-28 11:17:58 +01:00
Florian Kaltenberger 031535e95c Merge branch 'develop' into bugfix-40-channel_bandwidth 2015-10-26 18:28:48 +01:00
Florian Kaltenberger fb902d9871 Merge branch 'develop' into bugfix-40-channel_bandwidth 2015-10-26 18:26:53 +01:00