search for: restore_fpu

...st,int _dst_ystride, + const unsigned char *_src1,int _src1_ystride,const unsigned char *_src2, + int _src2_ystride,const ogg_int16_t *_residue); +void (* oc_state_frag_copy) (const struct oc_theora_state *_state, const int *_fragis, + int _nfragis,int _dst_frame,int _src_frame,int _pli); +void (* restore_fpu)(void); +} oc_base_opt_vtable; /*Common state information between the encoder and decoder.*/ @@ -291,10 +306,12 @@ oc_quant_tables dequant_table_data[2][3]; /*Loop filter strength parameters.*/ unsigned char loop_filter_limits[64]; + /*Table for accelerated common funct...

Hi all, I've patched theora-mmx to build on x86_64. The patch against SVN is attached. Basically all I did was to copy lib/i386 to lib/x86_64 and tweak the assembler code a bit: * added to each file: typedef unsigned long int ogg_uint64_t * converted all asm inputs to 64-bit in: dsp_mmxext.c, fdct_mmx.c, recon_mmx.c * left all asm outputs at 32-bit * I didn't patch dsp_mmx.c, since all

...st,int _dst_ystride, + const unsigned char *_src1,int _src1_ystride,const unsigned char *_src2, + int _src2_ystride,const ogg_int16_t *_residue); +void (* oc_state_frag_copy) (const struct oc_theora_state *_state, const int *_fragis, + int _nfragis,int _dst_frame,int _src_frame,int _pli); +void (* restore_fpu)(void); +} oc_base_opt_vtable; /*Common state information between the encoder and decoder.*/ @@ -291,10 +306,12 @@ oc_quant_tables dequant_table_data[2][3]; /*Loop filter strength parameters.*/ unsigned char loop_filter_limits[64]; + /*Table for accelerated common funct...

...rnal.h b/lib/internal.h --- a/lib/internal.h 2005-08-17 09:29:57.318872500 +0200 +++ b/lib/internal.h 2005-08-17 10:05:34.076411500 +0200 @@ -242,6 +242,8 @@ int _pli,ogg_int16_t _dct_coeffs[64],int _last_zzi,int _ncoefs, ogg_uint16_t _dc_iquant,const ogg_uint16_t _ac_iquant[64]); void (*restore_fpu)(void); + void (*oc_state_loop_filter_frag_rows)(oc_theora_state *_state,int *_bv, + int _refi,int _pli,int _fragy0,int _fragy_end); }oc_base_opt_vtable; @@ -391,8 +393,6 @@ int oc_state_loop_filter_init(oc_theora_state *_state,int *_bv); void oc_state_loop_filter(oc_theora_state *_stat...

...l/traps-xen.c @@ -1057,6 +1057,7 @@ asmlinkage void math_state_restore(struc struct task_struct *tsk = thread->task; /* NB. ''clts'' is done for us by Xen during virtual trap. */ + __get_cpu_var(xen_x86_cr0) &= ~X86_CR0_TS; if (!tsk_used_math(tsk)) init_fpu(tsk); restore_fpu(tsk); --- a/arch/x86_64/kernel/process-xen.c +++ b/arch/x86_64/kernel/process-xen.c @@ -574,6 +574,8 @@ __switch_to(struct task_struct *prev_p, BUG_ON(mcl > _mcl + ARRAY_SIZE(_mcl)); if (unlikely(HYPERVISOR_multicall_check(_mcl, mcl - _mcl, NULL))) BUG(); + if (_mcl->op == __HYPERVISO...

...int _pli); void (*state_frag_recon)(oc_theora_state *_state,const oc_fragment *_frag, - int _pli,ogg_int16_t _dct_coeffs[64],int _last_zzi,int _ncoefs, + int _pli,ogg_int16_t _dct_coeffs[128],int _last_zzi,int _ncoefs, ogg_uint16_t _dc_iquant,const ogg_uint16_t _ac_iquant[64]); void (*restore_fpu)(void); void (*oc_state_loop_filter_frag_rows)(oc_theora_state *_state,int *_bv, @@ -409,7 +409,7 @@ void oc_state_frag_copy(const oc_theora_state *_state,const int *_fragis, int _nfragis,int _dst_frame,int _src_frame,int _pli); void oc_state_frag_recon(oc_theora_state *_state,const oc_fragm...

quite some speed improvement indeed. attached the updated patch to apply to svn/trunk. j -------------- next part -------------- A non-text attachment was scrubbed... Name: theora-mmx.patch.gz Type: application/x-gzip Size: 8648 bytes Desc: not available Url : http://lists.xiph.org/pipermail/theora-dev/attachments/20040824/5a5f2731/theora-mmx.patch-0001.bin

...0.2499 382 0.0984 dump oc_token_skip_run 1275 0.2258 1598 0.4117 dump oc_dec_mb_modes_unpack 1008 0.1785 460 0.1185 dump oc_token_expand_cati 999 0.1769 1538 0.3963 dump oc_restore_fpu 884 0.1565 581 0.1497 dump oc_token_dec1val_const 614 0.1087 145 0.0374 dump oc_token_expand_zrl 543 0.0962 1 2.6e-04 dump oc_frag_recon_inter 509 0.0901 15 0.0039 dump...