From cd577ecd10b36ff15e5904ca3dbd43413d907543 Mon Sep 17 00:00:00 2001 From: gabest11 Date: Sun, 22 Jan 2012 13:00:45 +0000 Subject: [PATCH] GSdx: a different fix for ZoE2, also seem to help DMC, so I removed the previous one, please check again. git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5080 96395faa-99c1-11dd-bbfe-3dabce05a288 --- plugins/GSdx/GSRendererSW.cpp | 52 +++++++++++++++++----------------- plugins/GSdx/GSVertexTrace.cpp | 29 +++++++++++++++---- 2 files changed, 50 insertions(+), 31 deletions(-) diff --git a/plugins/GSdx/GSRendererSW.cpp b/plugins/GSdx/GSRendererSW.cpp index dcdf869070..7ce9487b21 100644 --- a/plugins/GSdx/GSRendererSW.cpp +++ b/plugins/GSdx/GSRendererSW.cpp @@ -462,13 +462,21 @@ void GSRendererSW::Sync(int reason) { s_n++; - std::string s = format("c:\\temp1\\_%05d_f%lld_rt1_%05x_%d.bmp", s_n, m_perfmon.GetFrame(), m_context->FRAME.Block(), m_context->FRAME.PSM); + std::string s; + + if(s_save) + { + s = format("c:\\temp1\\_%05d_f%lld_rt1_%05x_%d.bmp", s_n, m_perfmon.GetFrame(), m_context->FRAME.Block(), m_context->FRAME.PSM); - m_mem.SaveBMP(s, m_context->FRAME.Block(), m_context->FRAME.FBW, m_context->FRAME.PSM, GetFrameRect().width(), 512); + m_mem.SaveBMP(s, m_context->FRAME.Block(), m_context->FRAME.FBW, m_context->FRAME.PSM, GetFrameRect().width(), 512); + } - s = format("c:\\temp1\\_%05d_f%lld_zb1_%05x_%d.bmp", s_n, m_perfmon.GetFrame(), m_context->ZBUF.Block(), m_context->ZBUF.PSM); + if(s_savez) + { + s = format("c:\\temp1\\_%05d_f%lld_zb1_%05x_%d.bmp", s_n, m_perfmon.GetFrame(), m_context->ZBUF.Block(), m_context->ZBUF.PSM); - m_mem.SaveBMP(s, m_context->ZBUF.Block(), m_context->FRAME.FBW, m_context->ZBUF.PSM, GetFrameRect().width(), 512); + m_mem.SaveBMP(s, m_context->ZBUF.Block(), m_context->FRAME.FBW, m_context->ZBUF.PSM, GetFrameRect().width(), 512); + } } t = __rdtsc() - t; @@ -490,6 +498,8 @@ void GSRendererSW::InvalidateVideoMem(const GIFRegBITBLTBUF& BITBLTBUF, const GS if(!m_rl->IsSynced()) { + if(LOG) {fprintf(s_fp, "w %05x %d %d\n", BITBLTBUF.DBP, BITBLTBUF.DBW, BITBLTBUF.DPSM); fflush(s_fp);} + for(uint32* RESTRICT p = m_tmp_pages; *p != GSOffset::EOP; p++) { if(m_fzb_pages[*p] | m_tex_pages[*p]) @@ -508,6 +518,8 @@ void GSRendererSW::InvalidateLocalMem(const GIFRegBITBLTBUF& BITBLTBUF, const GS { if(!m_rl->IsSynced()) { + if(LOG) {fprintf(s_fp, "r %05x %d %d\n", BITBLTBUF.SBP, BITBLTBUF.SBW, BITBLTBUF.SPSM); fflush(s_fp);} + GSOffset* o = m_mem.GetOffset(BITBLTBUF.SBP, BITBLTBUF.SBW, BITBLTBUF.SPSM); o->GetPages(r, m_tmp_pages); @@ -575,12 +587,12 @@ void GSRendererSW::CheckDependencies(SharedData* sd) uint32* fb_pages = NULL; uint32* zb_pages = NULL; - if(sd->global.sel.fwrite) + if(sd->global.sel.fb) { fb_pages = m_context->offset.fb->GetPages(r); } - if(sd->global.sel.zwrite) + if(sd->global.sel.zb) { zb_pages = m_context->offset.zb->GetPages(r); } @@ -637,8 +649,8 @@ bool GSRendererSW::CheckTargetPages(const uint32* fb_pages, const uint32* zb_pag { bool synced = m_rl->IsSynced(); - bool fwrite = fb_pages != NULL; - bool zwrite = zb_pages != NULL; + bool fb = fb_pages != NULL; + bool zb = zb_pages != NULL; if(m_fzb != m_context->offset.fzb4) { @@ -755,7 +767,7 @@ bool GSRendererSW::CheckTargetPages(const uint32* fb_pages, const uint32* zb_pag // chross-check frame and z-buffer pages, they cannot overlap with eachother and with previous batches in queue, // have to be careful when the two buffers are mutually enabled/disabled and alternating (Bully FBP/ZBP = 0x2300) - if(fwrite) + if(fb) { for(const uint32* p = fb_pages; *p != GSOffset::EOP; p++) { @@ -768,7 +780,7 @@ bool GSRendererSW::CheckTargetPages(const uint32* fb_pages, const uint32* zb_pag } } - if(zwrite) + if(zb) { for(const uint32* p = zb_pages; *p != GSOffset::EOP; p++) { @@ -783,18 +795,6 @@ bool GSRendererSW::CheckTargetPages(const uint32* fb_pages, const uint32* zb_pag } } - if(!synced) - { - // FIXME: DMC3 (FBP == ZBP == 0x01000) - - if(fwrite && zwrite && m_context->FRAME.FBP == m_context->ZBUF.ZBP) - { - if(LOG) {fprintf(s_fp, "syncpoint 4\n"); fflush(s_fp);} - - return true; - } - } - return false; } @@ -1324,12 +1324,12 @@ GSRendererSW::SharedData::~SharedData() { if(m_using_pages) { - if(global.sel.fwrite) + if(global.sel.fb) { m_parent->ReleasePages(m_fb_pages, 0); } - if(global.sel.zwrite) + if(global.sel.zb) { m_parent->ReleasePages(m_zb_pages, 1); } @@ -1354,12 +1354,12 @@ void GSRendererSW::SharedData::UseTargetPages(const uint32* fb_pages, const uint m_fb_pages = fb_pages; m_zb_pages = zb_pages; - if(global.sel.fwrite) + if(global.sel.fb) { m_parent->UsePages(fb_pages, 0); } - if(global.sel.zwrite) + if(global.sel.zb) { m_parent->UsePages(zb_pages, 1); } diff --git a/plugins/GSdx/GSVertexTrace.cpp b/plugins/GSdx/GSVertexTrace.cpp index 1d1ea056c9..1d9fb53641 100644 --- a/plugins/GSdx/GSVertexTrace.cpp +++ b/plugins/GSdx/GSVertexTrace.cpp @@ -133,14 +133,22 @@ void GSVertexTrace::FindMinMax(const void* vertex, const uint32* index, int coun break; } - GSVector4 pmin = s_minmax.xxxx(); - GSVector4 pmax = s_minmax.yyyy(); GSVector4 tmin = s_minmax.xxxx(); GSVector4 tmax = s_minmax.yyyy(); GSVector4i cmin = GSVector4i::xffffffff(); GSVector4i cmax = GSVector4i::zero(); - // TODO: SSE41 has integer min/max, use that for xy/z/uv/f + #if _M_SSE >= 0x401 + + GSVector4i pmin = GSVector4i::xffffffff(); + GSVector4i pmax = GSVector4i::zero(); + + #else + + GSVector4 pmin = s_minmax.xxxx(); + GSVector4 pmax = s_minmax.yyyy(); + + #endif const GSVertex* RESTRICT v = (GSVertex*)vertex; @@ -198,18 +206,29 @@ void GSVertexTrace::FindMinMax(const void* vertex, const uint32* index, int coun GSVector4i xy = xyzf.upl16(); GSVector4i z = xyzf.yyyy().srl32(1); + #if _M_SSE >= 0x401 + + GSVector4i p = xy.blend16<0xf0>(z.uph32(!sprite ? xyzf : f)); + + pmin = pmin.min_u32(p); + pmax = pmax.max_u32(p); + + #else + GSVector4 p = GSVector4(xy.upl64(z.upl32(!sprite ? xyzf.wwww() : f))); pmin = pmin.min(p); pmax = pmax.max(p); + + #endif } } GSVector4 o(context->XYOFFSET); GSVector4 s(1.0f / 16, 1.0f / 16, 2.0f, 1.0f); - m_min.p = (pmin - o) * s; - m_max.p = (pmax - o) * s; + m_min.p = (GSVector4(pmin) - o) * s; + m_max.p = (GSVector4(pmax) - o) * s; if(tme) {